MILD Combustion of Methanol,Ethanol and 1-Butanol binary blends with Ammonia

In the present energy transition scenario, ammonia is considered a valuable candidate as energy-dense carrier with neutral or even negative carbon balance. However, the potential high NO_x emissions and the reduced oxidation process stability, at least when conventional combustion plants are used, can burden its wide utilization on large scales. In this context, MILD Combustion, due to its inherent characteristics, may greatly improve combustion stability and keep the NO_x emissions at an acceptable level. On the other hand, the addition of low or no-carbon fuels from biomasses and wastes, more reactive than ammonia, may be beneficial in further improving its combustion performance and the global sustainability of the energy supply chain.In this respect, the present work analyzes the sustainability and combustion performance of binary mixtures of ammonia and low-molecular-weight alcohols in a cyclonic burner, where MILD conditions are attained by means of a strong internal recirculation, and compares them with those obtained with NH₃/methane blends. Results highlighted that NH₃/alcohols mixtures ensure a stable oxidation process in a wide range of operational parameters without compromising the system performance. Moreover, they showed a significant reduction of NO_x emissions for NH₃/alcohols mixtures, especially for fuel-lean conditions, when compared to NH₃/methane blends.Experimental data were also corroborated by chemical kinetic modeling results to provide some insights on the peculiar NO_x formation routes when blends of different nature are used, highlighting the interaction between carbon and nitrogen fuels kinetics.     

Fuel and thermal load flexibility of a MILD burner

The strategy to stabilize distributed combustion regimes adopting cyclonic flow fields has been proven to be challenging. In fact, the establishment of a toroidal flow field within a combustion chamber may ensure the recirculation of mass and sensible enthalpy required to simultaneously dilute the fresh reactants and increase the temperature above the autoignition one. The combination of reactants dilution and preheating may greatly increase system energy efficiency and lower pollutants production producing very peculiar combustion regime (named MILD Combustion). At the same time this strategy can be compromised if the sensible enthalpy is not high enough to promote the auto-ignition process of diluted mixtures. This may happen either because of an inefficient recirculation system or due to a too low calorific fuel. The paper aims at exploiting the performance of a small-size cyclonic burner for a conventional fuel (CH₄) and a low calorific fuel (synthetic biogas) through the characterization of the process stabilization and pollutant emissions as a function of the mixture equivalence ratio and the nominal thermal power of the inlet mixture (from 2 to 10?kW), with the aim of identifying the optimal operating condition of the system. Results suggest that the system has to be exercised with mixtures with compositions slightly under the stoichiometric conditions and in a well identified temperature range to minimize both NO_x and CO emission. The burner can be easily exercised also with low calorific fuels for higher thermal powers according to the low LHV. However, it results that an efficient recirculation of the exhausts produces a robust MILD combustion condition also when low calorific fuels are used.     

Experimental study on co-firing characteristics of ammonia with pulverized coal in a staged combustion drop tube furnace

Utilizing ammonia as a co-firing fuel to replace amounts of fossil fuel seems a feasible solution to reduce carbon emissions in existing pulverized coal-fired power plants. However, there are some problems needed to be considered when treating ammonia as a fuel, such as low flame stability, low combustion efficiency, and high NO_x emission. In this study, the co-firing characteristics of ammonia with pulverized coal are studied in a drop tube furnace with staged combustion strategy. Results showed that staged combustion would play a key role in reducing NO_x emissions by reducing the production of char-NO_x and fuel(NH₃)-NO_x simultaneously. Furthermore, the effects of different ammonia co-firing methods on the flue gas properties and unburned carbon contents were compared to achieve both efficient combustion and low NO_x emission. It was found that when ammonia was injected into 300 mm downstream under the condition of 20% co-firing, lower NO_x emission and unburnt carbon content than those of pure coal combustion can be achieved. This is probably caused by a combined effect of a high local equivalence ratio of NH₃/air and the prominent denitration effect of NH₃ in the vicinity of the NH₃ downstream injection location. In addition, NO_x emissions can be kept at approximately the same level as coal combustion when the co-firing ratio is below 30%. And the influence of reaction temperature on NO_x emissions is closely associated with the denitration efficiency of the NH₃. Almost no ammonia slip has been detected for any injection methods and co-firing ratio in the studied conditions. Thus, it can be confirmed that ammonia can be used as an alternative fuel to realize CO₂ reduction without extensive retrofitting works. And the NO_x emission can be reduced by producing a locally NH₃ flame zone with a high equivalence ratio as well as ensuring adequate residence time.     

Towards the development of an efficient low-NOx ammonia combustor for a micro gas turbine

Recent studies have demonstrated stable generation of power from pure ammonia combustion in a micro gas turbine (MGT) with a high combustion efficiency, thus overcoming some of the challenges that discouraged such applications of ammonia in the past. However, achievement of low NOx emission from ammonia combustors remains an important challenge. In this study, combustion techniques and combustor design for efficient combustion and low NOx emission from an ammonia MGT swirl combustor are proposed. The effects of fuel injection angle, combustor inlet temperature, equivalence ratio, and ambient pressure on flame stabilization and emissions were investigated in a laboratory high pressure combustion chamber. An FTIR gas analyser was employed in analysing the exhaust gases. Numerical modeling using OpenFOAM was done to better understand the dependence of NO emissions on the equivalence ratio. The result show that inclined fuel injection as opposed to vertical injection along the combustor central axis resulted to improved flame stability, and lower NH₃ and NOx emissions. Numerical and experimental results showed that a control of the equivalence ratio upstream of the combustor is critical for low NOx emission in a rich-lean ammonia combustor. NO emission had a minimum value at an upstream equivalence ratio of 1.10 in the experiments. Furthermore, NO emission was found to decrease with ambient pressure, especially for premixed combustion. For the rich-lean combustion strategy employed in this study, lower NOx emission was recorded in premixed combustion than in non-premixed combustion indicating the importance of mixture uniformity for low NOx emission from ammonia combustion. A prototype liner developed to enhance the control and uniformity of the equivalence ratio upstream of the combustor further improved ammonia combustion. With the proposed liner design, NOx emission of 42?ppmv and ammonia combustion efficiency of 99.5% were achieved at 0.3?MPa for fuel input power of 31.44?kW.     

Simulations of laminar non-premixed flames of kerosene with hot combustion products as oxidiser

The effects of hot combustion product dilution in a pressurised kerosene-burning system at gas turbine conditions were investigated with laminar counterflow flame simulations. Hot combustion products from a lean (φ = 0.6) premixed flame were used as an oxidiser with kerosene surrogate as fuel in a non-premixed counterflow flame at 5, 7, 9 and 11 bar. Kerosene-hot product flames, referred to as ‘MILD’, exhibit a flame structure similar to that of kerosene–air flames, referred to as ‘conventional’, at low strain rates. The Heat Release Rate (HRR) of both conventional and MILD flames reflects the pyrolysis of the primary and intermediate fuels on the rich side of the reaction zone. Positive HRR and OH regions in mixture fraction space are of similar width to conventional kerosene flames, suggesting that MILD flames are thin fronts. MILD flames do not exhibit typical extinction behaviour, but gradually transition to a mixing solution at very high rates of strain (above A = 160, 000 s^?1 for all pressures). This is in agreement with literature that suggests heavily preheated and diluted flames have a monotonic S-shaped curve. Despite these differences in comparison with kerosene–air flames, MILD flames follow typical trends as a function of both strain and pressure. Further still, the peak locations of the overlap of OH and CH₂O mass fractions in comparison with the peak HRR indicate that the pixel-by-pixel product of OH- and CH₂O-PLIF signals is a valid experimental marker for non-premixed kerosene MILD and conventional flames.     

Stability and characteristics of NH3/CH4/air flames in a combustor fired by a double swirl stabilized burner

Compared to hydrocarbons, ammonia's low reactivity and higher NO_x emissions limit its practical application. Consequently, its implementation in combustion systems requires a different combustor geometry, by adapting existing systems or developing new ones. This study investigates the flame stability, NO emissions, and flame structure of NH₃/CH₄/air premixed flames in a novel combustor comprising a double swirl burner. A lean premixed CH₄/air mixture of equivalence ratio, Φ_out, was supplied to the outer swirl, while a NH₃/CH₄/Air mixture fed the inner swirl. The molar fraction of NH₃ in the inner fuel blend, x_NH3, was varied from 0 (pure CH₄) to 1 (pure NH₃) over far-lean to far-rich inner stream equivalence ratio, Φ_in. This new burner's stability map was established in terms of Φ_in versus x_NH3 for different Φ_out. Then, NO emissions were measured versus Φ_in for various x_NH3 and Φ_out. Finally, based on the NO emissions, eight flames were down-selected for in-flame measurements, which included temperature and OH-PLIF. The stability measurements revealed that increasing x_NH3 modifies the stability map by increasing the lean blowout limits and narrowing the flashback region. At Φ_out ≥ 0.6, a stable flame was achieved for a pure inner NH₃/air mixture. Low NO emissions were achieved in this burner configuration at x_NH3=1 by either enriching or far-leaning Φ_in. Enriching Φ_in led to a steep decrease in NO concentrations. However, to achieve low NO concentrations, precise control of Φ_out was needed. At Φ_in=1.4, 220 ppm NO at Φ_out=0.7 versus 690 at Φ_out=0.6 was measured. Moreover, substantially enriching Φ_in>1.2 led to a slight decrease in measured NO. Generally, the OH-PLIF images revealed a conical OH-layer at the burner exit. Certain flame conditions created OH-pockets inside the conical structure or formed a V-shaped OH-layer far downstream. This change in flame structure was found to impact NO emissions strongly.     

Transported PDF simulation of turbulent CH4/H2 flames under MILD conditions with particle-level sensitivity analysis

Transported probability density function (TPDF) simulation with sensitivity analysis has been conducted for turbulent non-premixed CH₄/H₂ flames of the jet-into-hot-coflow (JHC) burner, which is a typical model to emulate moderate or intense low oxygen dilution combustion (MILD). Specifically, two cases with different levels of oxygen in the coflow stream, namely HM1 and HM3, are simulated to reveal the differences between MILD and hot-temperature combustion. The TPDF simulation well predicts the temperature and species distributions including those of OH, CO and NO for both cases with a 25-species mechanism. The reduced reaction activity in HM1 as reflected in the peak OH concentration is well correlated to the reduced oxygen in the coflow stream. The particle-level local sensitivities with respect to mixing and chemical reaction further show dramatic differences in the flame characteristics. HM1 is less sensitive to mixing and reaction parameters than HM3 due to the suppressed combustion process. Specifically, for HM1 the sensitivities to mixing and chemical reactions have comparable magnitude, indicating that the combustion progress is controlled by both mixing and reaction in MILD combustion. For HM3, there is however a change in the combustion mode: during the flame initialization, the combustion progress is more sensitive to chemical reactions, indicating that finite-rate chemistry is the controlling process during the autoignition process for flame stabilization; at further downstream where the flame has established, the combustion progress is controlled by mixing, which is characteristic of nonpremixed flames. An examination of the particles with the largest sensitivities reveals the difference in the controlling mixtures for flame stabilization, namely, the stoichiometric mixtures are important for HM1, whereas, fuel-lean mixtures are controlling for HM3. The study demonstrates the potential of TPDF simulations with sensitivity analysis to investigate the effects of finite-rate chemistry on the flame characteristics and emissions, and reveal the controlling physio-chemical processes in MILD combustion.     

Effects of residence time on the NOx emissions of premixed ammonia-methane-air swirling flames at elevated pressure

In this study, a bespoke single-stage swirl burner was used to experimentally investigate the effects of residence time on emissions from premixed ammonia-methane-air flames. The residence time was altered in two ways: by modifying the combustion chamber's length or by modifying the swirl number. Exhaust emissions of O₂, CO₂, CO, NO, NO₂, and N₂O were measured at an absolute pressure of 2 bar for equivalence ratios between 0.50 and 0.95 and ammonia fractions in the fuel blend between 0 and 100%. Spatial distributions of NO and OH radicals were also imaged using PLIF inside the combustion chamber at different heights above the nozzle. Data shows that increasing residence time can further advance chemical reactions, as evidenced by a reduction in O₂ concentration in the exhaust. Increasing the swirl number reduces emissions of NO, NO₂, and N₂O more efficiently than tripling the chamber's length. However, a decrease in the combustion efficiency may be responsible for a fraction of this NOx reduction when the swirl number is increased for some equivalence ratios. NO emissions are not modified when the chamber's length is increased, which is consistent with the fact that the NO-LIF signal does not decay when the distance from the nozzle increases. Therefore, NO formation is somehow restricted to within the main reaction zone of the swirling flame, that is, the zone whose height does not exceed 60 mm for this burner. Conversely, tripling the chamber's length reduces the concentrations of NO₂ and N₂O. This reduction is not reflected in a measurable increase in NO concentration because NO is present in much larger quantities than NO₂ and N₂O in flames examined here. Consistent with the fact that OH promotes NO formation via fuel-NOx pathways, a positive correlation is found between NO- and OH-LIF intensities.     

SGS Reaction rate modelling for MILD combustion based on machine-learning combustion mode classification: Development and a priori study

A neural network (NN) aided model is proposed for the filtered reaction rate in moderate or intense low-oxygen dilution (MILD) combustion. The framework of the present model is based on the partially stirred reactor (PaSR) approach, and the fraction of the reactive structure appearing in the PaSR is predicted using different NN’s, to consider both premixed and non-premixed conditions while allowing the use of imbalanced training data between premixed and non-premixed combustion direct numerical simulation (DNS) data. The key ingredient in the present model is the use of local combustion mode prediction performed by using another NN, which is developed in a previous study. The trained model was then assessed by using two unknown combustion DNS cases, which yields much higher dilution level (more intense MILD condition) and higher Karlovitz number than the DNS cases used as training data. The model performance assessment has been carried out by means of the Pearson’s correlation coefficient and mean squared error. For both the present model and zeroth-order approximated reaction rate, the correlation coefficient with the target values shows relatively high values, suggesting that the trend of predicted field, by the present model and zeroth-order approximation, is well correlated with the actual reaction rate field. This suggests that the use of PaSR equation is promising if the fraction of the reactive structure is appropriately predicted, which is the objective in the present study. On the other hand, substantially lower mean squared error is observed for a range of filter sizes for the present model than that for the zeroth-order approximation. This suggests that the present filtered reaction rate model can account for the SGS contribution reasonably well.     

Identification of combustion mode under MILD conditions using Chemical Explosive Mode Analysis

Direct Numerical Simulations (DNS) data of Moderate or Intense Low-oxygen Dilution (MILD) combustion are analysed to identify the contributions of the autoignition and flame modes. This is performed using an extended Chemical Explosive Mode Analysis (CEMA) which accounts for diffusion effects allowing it to discriminate between deflagration and autoignition. This analysis indicates that in premixed MILD combustion conditions, the main combustion mode is ignition for all dilution and turbulence levels and for the two reactant temperature conditions considered. In non-premixed conditions, the preponderance of the ignition mode was observed to depend on the axial location and mixture fraction stratification. With a large mixture fraction lengthscale, ignition is more preponderant in the early part of the domain while the deflagrative mode increases further downstream. On the other hand, when the mixture fraction lengthscale is small, sequential autoignition is observed. Finally, the various combustion modes are observed to correlate strongly with mixture fraction where lean mixtures are more likely to autoignite while stoichiometric and rich mixtures are more likely to react as deflagrative structures.     

Effects of radiation heat loss on laminar premixed ammonia/air flames

Ammonia (NH₃) direct combustion is attracting attention for energy utilization without CO₂ emissions, but fundamental knowledge related to ammonia combustion is still insufficient. This study was designed to examine effects of radiation heat loss on laminar ammonia/air premixed flames because of their very low flame speeds. After numerical simulations for 1-D planar flames with and without radiation heat loss modeled by the optically thin model were conducted, effects of radiation heat loss on flame speeds, flame structure and emissions were investigated. Simulations were also conducted for methane/air mixtures as a reference. Effects of radiation heat loss on flame speeds were strong only near the flammability limits for methane, but were strong over widely diverse equivalence ratios for ammonia. The lower radiative flame temperature suppressed the thermal decomposition of unburned ammonia to hydrogen (H₂) at rich conditions. The equivalence ratio for a low emission window of ammonia and nitric oxide (NO) in the radiative condition shifted to a lower value than that in the adiabatic condition.     

Numerical simulation of a mixed-mode reaction front in a PPC engine

The ignition process, mode of combustion and reaction front propagation in a partially premixed combustion (PPC) engine running with a primary reference fuel (87% iso-octane, 13% n-heptane by volume) is studied numerically in a large eddy simulation. Different combustion modes, ignition front propagation, premixed flame and non-premixed flame, are observed simultaneously. Displacement speed of CO iso-surface propagation describes the transition of premixed auto-ignition to non-premixed flame. High temporal resolution optical data of CH₂O and chemiluminescence are compared with simulated results. A high speed ignition front is seen to expand through fuel-rich mixture and stabilize around stoichiometry in a non-premixed flame while lean premixed combustion occurs in the spray wake at a much slower pace. A good qualitative agreement of the distribution of chemiluminescence and CH₂O formation and destruction shows that the simulation approach sufficiently captures the driving physics of mixed-mode combustion in PPC engines. The study shows that the transition from auto-ignition to flame occurs over a period of several crank angles and the reaction front propagation can be captured using the described model.     

Effects of CO2 dilution on partially premixed CH4-air flames in swirl and bluff body stabilized combustor

The effect of CO₂ dilution on the flame characteristics and pollutant emission of a partially premixed CH₄-air flame in a confined bluff body and swirl influenced flowfield is investigated using optical and laser diagnostic methods. The non-premixed burner produced a converging-diverging flowfield at the burner exit and a lifted flame is produced at all test cases, with an upstream movement of the flame with decreasing global equivalence ratios (?_g). Based on variations in ?_g, two flame stabilization modes – bluff body influenced and swirl stabilized – with a transition mode in-between is observed for the cases with (flame FB) and without dilution (flame FM). The characteristics of the heat release zone are influenced by dilution, with the FB flames being longer and also less intense when compared to FM flames. Pollutant measurement at 30 mm downstream from the combustor exit highlighted the ultra-low NO_x capability of the IIST-GS2 burner. CO₂ dilution leads to a reduction in NO_x emission due to both thermal and chemical effects. For ?_g ≥ 0.7 extreme low levels of CO and unburned hydrocarbons (UHC) are observed for both cases. For ?_g ≤ 0.6 the dramatic increase of both CO and UHC maybe due to the lower flame temperatures and shorter flame zone residence times, respectively.     

Experimental and numerical study of premixed, lean ethylene flames

Ethylene is a key intermediate in the combustion mechanisms of most practical fuels. It plays also an important role in the formation of aromatic hydrocarbons and soot particules. The latter has motivated many experimental and numerical studies carried out on rich ethylene-air mixtures. Less studies have been devoted to lean mixtures, and the development of strategies based on lean, premixed flames to reduce soot and NO_x production requires additional experimental data in lean conditions. In this work, the chemical structure of lean premixed ethylene-oxygen-nitrogen flames stabilized on a flat-flame burner at atmospheric pressure was determined experimentally. The species mole fraction profiles were also computed by the Premix code (Chemkin II version) and four detailed reaction mechanisms. A very good agreement was observed for the main flame properties: reactants consumption, final products (CO₂, H₂O) and the main intermediates: CO and H₂. Marked differences occurred in the prediction of active intermediate species present in small concentrations. Pathways analyses were performed to identify the origins of these discrepancies. It was shown that the same reactions were involved in the four mechanisms to describe the consumption of ethylene, but with marked differences in their relative importance. C₂H₃ and CH₂HCO are the main radicals formed in this first step and their consumption increases the differences between the mechanisms either by the use of different kinetic data for common reactions or by differences in the nature of the consumption reactions.     

Measurements and modelling of oxy-fuel coal combustion

Oxy-fuel combustion is one of the most promising technologies to isolate efficiently and economically CO₂ emissions in coal combustion for the ready carbon sequestration. The high proportions of both H₂O and CO₂ in the furnace have complex impacts on flame characteristics (ignition, burnout, and heat transfer), pollutant emissions (NO_x, SO_x, and particulate matter), and operational concerns (ash deposition, fouling/slagging). In contrast to the existing literature, this review focuses on fundamental studies on both diagnostics and modelling aspects of bench- or lab-scale oxy-fuel combustion and, particularly, gives attention to the correlations among combustion characteristics, pollutant formation, and operational ash concerns. First, the influences of temperature and species concentrations (e.g., O₂, H₂O) on coal ignition, volatile combustion and char burning processes, for air- and oxy-firing, are comparatively evaluated and modelled, on the basis of data from optically-accessible set-ups including flat-flame burner, drop-tube furnace, and down-fired furnace. Then, the correlations of combustion-generated particulate/NO_x emissions with changes of combustion characteristics in both air and oxy-fuel firing modes are summarized. Additionally, ash deposition propensity, as well as its relation to the formation of fine particulates (i.e. PM_0.2, PM₁ and PM₁₀), for both modes are overviewed. Finally, future research topics are discussed. Fundamental oxy-fuel combustion research may provide an ideal alternative for validating CFD simulations toward industrial applications.     

Multi-species detection in spray flames with tunable excimer lasers

Planar imaging with tunable excimer-laser sheet illumination is used to determine spatial distributions of different species in liquid-fuelled spray flames of commercial oil burning furnaces. Two burner configurations, which differ only in the fuel/air mixing devices, are investigated to understand why one configuration yields 30% less NO_x emission. Iso-octane and n-heptane fuels are used. To understand the origin for NO_x reduction spatial distributions of reactants (fuel, O₂), the reaction intermediate OH and the pollutant NO are recorded. OH and O₂ are measured by LIPF, NO by LIF. Fuel distributions are determined by another broad-band emission, whose origin is not yet identified. Both single shot and averaged distributions are recorded. The averaged distributions are extremely reproducible and depend sensitively on details of the burner geometry and the fuel/air mixing device. They can clearly be used to distinguish fine details in different injection systems. The spatial distribution of different species relative to each other yield considerable insight in the differences between the two combustion processes. On the basis of purely qualitative visualization it is possible to understand the origin for NO_x reduction: it results from faster injection of air in the one fuel/air mixing device.     

Characteristics and structure of inverse flames of natural gas

Characteristics and structure of nominally non-premixed flames of natural gas are investigated using a burner that employs simultaneously two distinct features: fuel and oxidiser direct injection, and inverse fuel and oxidiser delivery. At low exit velocities, the result is an inverse diffusion flame that has been noted in the past for its low NO_x emissions, soot luminosity, and narrow stability limits. The present study aimed at extending the burner operating range, and it demonstrated that the inverse flame exhibits a varying degree of partial premixing dependent on the discharge nozzle conditions and the ratio of inner air jet and outer fuel jet velocities. These two variables affect the flame length, temperature distributions, and stability limits. Temperature measurements and Schlieren visualisation show areas of enhanced turbulent mixing in the shear region and the presence of a well-mixed reaction zone on the flame centreline. This reaction zone is enveloped by an outer diffusion flame, yielding a unique double-flame structure. As the fuel–air equivalence ratio is decreasing with an increase in the inner jet velocity, the well-mixed reaction zone extends considerably. These findings suggest a method for establishing a flame of uniform high temperature by optimising the coaxial nozzle geometry and flow conditions. The normalised flame length is decreasing exponentially with the air/fuel velocity ratio. Measurements demonstrate that the inverse flame stability limits change qualitatively with varying degree of partial premixing. At the low premixing level, the flame blow-out is a function of the inner and outer jet velocities and the nozzle conditions. The flame blow-out at high degree of partial premixing occurs abruptly at a single value of the inner air jet velocity, regardless of the fuel jet velocity and almost independent of the discharge nozzle conditions.     

An algorithm for LES of premixed compressible flows using the Conditional Moment Closure model

A novel numerical method has been developed to couple a recent high order accurate fully compressible upwind method with the Conditional Moment Closure combustion model. The governing equations, turbulence modelling and numerical methods are presented in full. The new numerical method is validated against direct numerical simulation (DNS) data for a lean premixed methane slot burner. Although the modelling approaches are based on non-premixed flames and hence not expected to be valid for a wide range of premixed flames, the predicted flame is just 10% longer than that in the DNS and excellent agreement of mean mass fractions, conditional mass fractions and temperature is demonstrated. This new numerical method provides a very useful framework for future application of CMC to premixed as well as non-premixed combustion.     

Large eddy simulation of multi-regime combustion with a two-progress variable approach for carbon monoxide

Simulations of two cases in a novel multi-regime burner configuration are undertaken using a presumed joint probability density function (PDF) approach with tabulated chemistry. The flame conditions are varied by changing the central jet equivalence ratio, which produces different multi-regime combustion modes in the non-premixed inner flame. An outer premixed flame and recirculation zone behind a bluff body are present to supply heat and combustion products to stabilise the inner flame. A two-progress variable approach is tested to improve predictions of carbon monoxide (CO) in the post-flame regions, where CO oxidation occurs. The large eddy simulation set-up and sub-grid combustion model are assessed through comparisons with time-averaged measurements for radial profiles at different streamwise locations. The jet break-up length, the shear layers and the mixture fraction distribution are well captured in both cases. The temperature distribution is well captured for the inner flame in each case but the temperature and mixture fraction are over predicted in the downstream regions of the outer premixed flame, which is due to increased dilatation that suppresses air entrainment. Improved predictions of the CO mass fraction are obtained for the outer premixed flames with the two-progress variable approach. Over predictions are seen in the upstream regions of the inner flame when the CO mass fraction is obtained from a look-up table, suggesting that the CO mass fraction should be transported to include the convection/diffusion balance in regions where there is no flame. Furthermore, transporting the CO mass fraction with a one-progress variable approach produces over predictions in the burnt regions, suggesting a two-progress variable model is needed to capture the consumption region of CO. The multi-regime combustion characteristics are observed to be stronger in flame MRB26b, where non-premixed and rich premixed combustion is present. For flame MRB18b, the non-premixed contribution is smaller and weak stratified combustion is observed.