Experimental and kinetic study on laminar flame speeds of ammonia/syngas/air at a high temperature and elevated pressure

Abstract

The laminar flame speeds of ammonia mixed with syngas at a high pressure, temperature, and different syngas ratios were measured. The data obtained were fitted at different pressures, temperatures, syngas ratios, and equivalence ratios. Four kinetic models (the Glarborg model, Shrestha model, Mei model, and Han model) were compared and validated with experimental data. Pathway, sensitivity and radical pool analysis are conducted to find out the deep kinetic insight on ammonia oxidation and NO formation. The pathway analysis shows that H abstraction reactions and NH_i combination reactions play important roles in ammonia oxidation. NO formation is closely related to H, OH, the O radical produced, and formation reactions. NO is mainly formed from reaction, HNO+ H= NO+ H₂. Furthermore, both ammonia oxidation and NO formation are sensitive to small radical reactions and ammonia related reactions.

Graphical abstract

Keywords

ammonia mixed with syngas / laminar flame speed / kinetic model / sensitivity analysis / pathway analysis

1 Introduction

Ammonia (NH₃) is now attracting attention due to its carbon-free property and is regarded as a hydrogen carrier [1]. With a relatively higher boiling temperature (–33.4°C) compared with hydrogen, ammonia has a great advantage in storage and transport. In addition, it is cost-effective and can be synthesized in a number of methods. These advantages make ammonia a possible alternative fuel for gas turbines, internal combustion engines, and space propulsion systems [2,3]. However, the high ignition energy, the low flame speeds, and the high fuel NO_x production [3,4] of ammonia inhibit the use of NH₃ as a fuel. Mei et al. [5] have revealed that oxygen enrichment is also a feasible. Therefore, the strategy of co-firing NH₃ with methane [6,7], hydrogen [8], dimethyl ether [9], gasoline, diesel etc. [10], has been presented. Among these candidates, syngas is a good choice, which consists of hydrogen and carbon-monoxide and has been proved to have low emissions and a high energy efficiency in many combustion processes such as the integrated gasification combined cycle (IGCC) [11]. Furthermore, the combustion kinetic and mechanism of syngas have been much more extensively studied than larger molecules. Co-firing with syngas can provide low-carbon fuel with a higher flame speed and a lower ignition energy than pure ammonia [1,12]. Considering the above advantages, co-firing NH₃ with syngas would offer a potential for the improvement of ammonia combustion. Detailed kinetic analysis on co-firing of NH₃ and syngas would help control NO_x emissions.

Among all the fundamental property for combustion, laminar flame speed is one of the key parameters to describe flame structures, velocity, extinction limit, and flame stabilization. Several experiments have been conducted to measure the laminar flame speed of NH₃/air [13,14]. The experimental data are used for model developed and validation. The results show that the flame velocity of NH₃ is relatively low, which demonstrates the necessity of co-firing with syngas. As for NH₃/syngas/air, Han et al. [14] experimentally investigated the laminar flame speed at atmospheric pressure and 298 K using the heat flux method and helped improve the existing mechanism. Wang et al. [15] measured the laminar flame speeds of NH₃/syngas/air, NH₃/CO/air, and NH₃/H₂/air by using the heat flux method at pressures up to 5 atm and 298 K. Pressure exponent factors on laminar flame speeds were tested and detailed kinetic analysis were conducted. Mei et al. [16] measured the laminar flame speed of ammonia mixed with syngas using combustion bomb at 1–10 atm and 298 K. It shows that the chemical effect is more responsible for the enhanced laminar flame propagation of ammonia with syngas. In addition, the existing data were measured at 298 K, limiting to low temperature. Laminar flame speeds of NH₃/syngas/air under wider conditions should be provided to validate the mechanisms.

There are also several numerical mechanisms available which include ammonia and syngas, such as San Diego model [17], Okafor et al. [18], Zhang et al. [19], and Shrestha et al. [20]. However, the recently published mechanism presented discrepancies for experimental data. As presented in Han et al. [14], the Okafor model and San Diego model show the same tendency in predicting the NH₃/syngas/air flames, both of which underestimate the laminar flame speed under different conditions while the Zhang model always over-predictes the laminar flame speed of ammonia with syngas. The Shrestha model shows a good agreement with the experimental data under the fuel lean condition but there exists over-prediction at fuel rich side. In addition, Han et al. [21] presented a modified model which matches well with the experimental data. Glarborg et al. [22] emphasized nitrogen chemistry in combustion of light hydrocarbons and this model is validated over a wide range of experimental data. Shrestha et al. [23] proposed a modified model for oxidation of ammonia and ammonia/hydrogen blends, which has a good agreement with measurements under wide conditions. The simulated results using Mei model [16] also matched well with the measured laminar flame speed and ignition delay times. Therefore, in this work the Han model, the Glarborg model, the Mei model, and the Shrestha model are validated for laminar flame of ammonia/syngas blends.

In this work, investigation is conducted on the blending effects of syngas on the combustion characteristics of ammonia. Laminar flame speeds for NH₃/syngas/air are measured at 1–3 atm, equivalence ratios from 0.7 to 1.6, 298–445 K, and mole fraction of syngas (x_syngas) from 0.1 to 1.0 in ammonia-syngas mixtures. The recently published mechanisms are validated against the experimental data. In addition, differences on laminar flame speeds and NO formation prediction using mechanisms are investigated. The chemical effect of H₂ and CO additions on NO formation is also discussed.

2 Experimental and numerical approach

The constant volume combustion bomb used in this work to measure the laminar flame speed is described in Refs. [24–26]. It is cylindrical in shape, with a volume of 5.8 L and two optical windows located in both sides with diameters of 80 mm. Ammonia (purity 99.9%) mixed with hydrogen (purity 99.9%), carbon monoxide (purity 99.9%) and air were prepared in the chamber after evacuation. The partial pressures of the components were controlled according to the required pressures using a combination of high precision pressure sensor with an accuracy of ±1 kPa and a micro-adjustable valve. The chamber was heated using heating tape to the designed temperature which was detected by a K type thermocouple and controlled using a temperature controller. Flame propagation was observed and recorded using a high-speed digital camera at 10000 frames per second via the quartz windows. Good experimental repeatability can be observed during the measurements. Finally, in data processing, only the radius from 8 to 22 mm in pictures was used as shown in Fig. S1 in Electronic Supplementary Material (ESM). The laminar flame speeds, S_u, were obtained using the nonlinear method proposed by Kelly et al. [27] as

(1){\left({\displaystyle \frac{s_{\mathrm{b}}}{s_{\mathrm{u}}}}\right)}^{\mathrm{2}}\ln {\left({\displaystyle \frac{s_{\mathrm{b}}}{s_{\mathrm{u}}}}\right)}^{\mathrm{2}}=-\mathrm{2}{\displaystyle \frac{L_{\mathrm{b}}k}{s_{\mathrm{u}}}},

where S_b= dr/dt is the stretched flame propagation speed , L_b is the Markstein length, and the flame stretch rate, k, is represented as

(2)k=\mathrm{2}{S}_{\mathrm{b}}/r={\displaystyle \frac{\mathrm{1}}{A}}{\displaystyle \frac{\mathrm{d}A}{\mathrm{d}t}},

where r is the flame radius, A refers to flame area.

The nonlinear method on the extraction of the unstretched flame speeds was shown in Fig. S2 (ESM). The purple line is the nonlinear fitted Markstein lengthen which indicates the sensitivity of the flame speed on extraction. The laminar flame speed exists where extraction is zero as the blue star shows.

The uncertainty of measured speeds was associated with mixture preparation, pressure, and temperature. The uncertainty of temperature is estimated to be 3 K according to the thermalcouple uncertainty, while the uncertainty in pressure controlled by high precision pressure transmitter is evaluated to be 1 kPa. In conclusion, the evaluated value of laminar flame speed uncertainty is 0.2 to 2.1 cm/s and will be shown as error bars.

The experimental conditions in this study are present in Table 1. The initial temperature is up to 443 K, while the pressure varies from 1 to 3 atm, and the mole fraction of syngas in fuel mixture (syngas/(syngas+ NH₃)), x_syn, changes from 0.1 to 1. For the equivalence ratio, experiment is conducted from 0.7 to 1.6 while for x_syn = 0.1, only the laminar flame speeds at 0.9 to 1.1 are measured.

All simulations of laminar flame speed were performed using CHEMKIN-II. As to each case, at least 500 grid points were used with Curve and Grad set to 0.02. Besides, the multi-component transport and Soret effect [28] were taken into consideration.

The Mei model [16], the Shrestha model [23], the Glarborg model [22], and the Han model [14] were used to predict the laminar flame speeds, which were recently proposed and validated under wide conditions on laminar flame speed of ammonia/hydrogen.

3 Results and discussion

3.1 Experimental validation

A comparison of the experimental data in this work and those in Refs. [14] and [16] was conducted to validate the accuracy of the experimental system as shown in Fig. 1. A good agreement can be observed between the data of this work and those in Refs. [14] and [16]. The laminar flame speed is highly sensitive to the H atom and the adiabatic flame temperature [16,25,29]. Therefore, with the increase in the mole fraction of the syngas, the H₂ in the syngas providing extra H source enhances the burning velocity. Figure 1 also presents the simulated results using the three mechanisms. They show the same trend as the experimental data. Both the Glarborg and Shrestha models over predicted the experimental data. Of these models, the Han model exhibits the best agreement with the experimental data.

3.2 Experimental data

The laminar flame speed experiment was conducted at temperatures from 298 to 443 K and at pressures from 1 to 3 atm, with equivalence ratios from 0.7 to 1.6. Figure 2 presents the schlieren images of the spherically propaga-ting flames of ammonia/syngas/air mixtures at x_syn = 50%, φ = 0.8, and T = 298 K. A smooth flame front can be observed at 1–2 atm. With pressure increasing, there exists a cellular structure at the flame front. When the mixtures become much richer, the cellular structure disappears, as shown in Fig. 3, indicating that the flame tends to be stable.

Figure 4 demonstrates the measured Markstein length (L_b) versus the equivalence ratio at different initial temperatures, pressures, and syngas ratios. It can be seen that the Markstein length increases monotonously with the increase of equivalence ratio under all conditions, and the increase tends to be more remarkable under the fuel rich condition. This indicates that the flame front stability is enhanced with the increase of the equivalence ratio. With less syngas, the Markstein lengths increase abruptly with the equivalence ratio while the L_b-φ tends to be flat with larger syngas ratios. This phenomenon can be explained with the classical models of flame instability due to the effects of preferential diffusion proposed by Manton et al. [30] and Markstein [31]. The laminar premixed flames will tend to be unstable due to the effects of preferential diffusion under the conditions where the fast-diffusing component (H₂ in the present instance) is deficient (corresponding to fuel lean combustion in this study).

Measured laminar flame speeds can be fitted at different pressures, temperatures, syngas ratios, and equivalence ratios, as expressed in Eq. (3) [32].

(3)\begin{array}{l}{S}_{\mathrm{u}}=S_{\mathrm{u}}^{\mathrm{0}}\left(\varphi \right){\left(T_{\mathrm{u}}/T_{\mathrm{u}\mathrm{0}}\right)}^{\alpha \left(\varphi \right)}{\left(P_{\mathrm{u}}/P_{\mathrm{u}\mathrm{0}}\right)}^{\beta \left(\varphi \right)}{\left(x_{\mathrm{u}}/x_{\mathrm{u}\mathrm{0}}\right)}^{\gamma \left(\varphi \right)},\hfill \\ \hfill \end{array}

where S_{\mathrm{u}}represents laminar flame speed, u means unburned gas, S_{\mathrm{u}}^{\mathrm{0}} means referenced laminar flame speed, α, β, and γ represent the exponent index of temperature, pressure, and syngas ratio, respectively. Moreover, \varphi refers to equivalence ratio. The referenced initial temperature T_{\mathrm{u}\mathrm{0}} is 298 K, the syngas ratio is 0.5, and the referenced initial pressure P_{\mathrm{u}\mathrm{0}} is 1 atm. It can be derived as

(4)S_{\mathrm{u}}^{\mathrm{0}}\left(\varphi \right)=-\mathrm{137.21}+\mathrm{388.65}\varphi -\mathrm{284.68}{\varphi }^{\mathrm{2}}+\mathrm{63.38}{\varphi }^{\mathrm{3}},

(5)\alpha \left(\varphi \right)=\mathrm{2.70}-\mathrm{2.23}\varphi +\mathrm{1.57}{\varphi }^{\mathrm{2}}-\mathrm{0.33}{\varphi }^{\mathrm{3}},

(6)\beta \left(\varphi \right)=-\mathrm{1.70}+\mathrm{3.39}\varphi -\mathrm{2.73}{\varphi }^{\mathrm{2}}+\mathrm{0.71}{\varphi }^{\mathrm{3}},

(7)\gamma \left(\varphi \right)=\mathrm{7.94}-\mathrm{17.06}\varphi +\mathrm{14.45}{\varphi }^{\mathrm{2}}-\mathrm{3.48}{\varphi }^{\mathrm{3}}.

Figure 5 is the comparison of the experimental data and fitted results at different pressures and temperatures. The fitted laminar flame speeds match well with the experimental data at temperatures of 298 to 443 K and under pressures from 1 to 3 atm, with equivalence ratios of 0.7 to 1.6. It can also be observed that laminar flame speeds increase with the increase of the mole fraction of syngas. Besides, the addition of syngas shifts the peak to the fuel rich condition.

3.3 Comparison of three models

Figure 6 depicts the measured and predicted laminar flame speeds of NH₃/syngas/air flame, as a function of equivalence ratio. The Glarborg model [22] over-predicts the laminar flame speeds under all conditions especially under at pressures and temperatures. The Shrestha model gives a reasonable prediction at 298 K under majority conditions. But it over-predicts the laminar flame speeds at high temperatures. Both the Mei and the Han model perform generally well on predicting the experimental results. The Han model slightly under-predicts the laminar flame speeds at high pressure. Discrepancies between simulated results using two models are less than 1.5 cm/s.

3.4 Kinetic analysis of NH₃/syngas oxidation

Since the Han model reproduces the experimental data well, it is adopted to analyze the oxidation of NH₃/syngas/air. This section investigates the effect of pressure, temperature, syngas ratio on the pathway, radical pools, and NO formation.

Figure 7 exhibits the reaction pathways at φ = 0.9 and 1.5 to investigate the effect of equivalence ratio on ammonia oxidation. Branching ratios in mole fraction are adopted for pathway analysis. Under the fuel lean condition, ammonia is totally consumed by H-abstraction reactions with O/H radicals, then producing NH₂. First, the H abstraction from NH₂ by OH and O radicals possesses the largest chain branching ratio of 59.1%. The second important pathways are combination reactions which consume 18.9% of NH₂. Finally, 14.8% of NH₂ is oxidized to HNO. Under the fuel rich condition, as the red tags show, the radical combination reactions become much more important than those in the fuel lean condition. The chain branching ratio of this reaction class increases to 40.9% with more fuel. In contrast, oxidation of NH₂ forming HNO can be neglected.

Furthermore, under the fuel lean condition, 47.2% of HNO is oxidized to NO through reactions HNO+ H= NO+ H₂ and HNO+ OH= NO+ H₂O, which is the most significant way for NO formation. NH and N radicals can also be transformed to NO through reactions NH+ O= NO+ H and N+ OH= NO+ H. Under the fuel rich condition, with the short of H, OH, and O radicals, less HNO is produced and NO formation mainly comes from the further oxidation of the N atom. In short, NO formation is closely related to OH, H, O, NH₂, NH, and N radicals. The majority of NO can be regarded as fuel NO as N mainly comes from ammonia. Therefore, NO formation poses a great challenge to the application of ammonia as fuel. It is also worth mentioning that the main consumption channels of NO are reactions with NH_i, leading to the formation of N₂O and N₂.

At φ = 1.5, combination of NH_i radicals possesses larger branching ratios, while the oxidation of NH and NH₂ forming HNO becomes less important. Moreover, chain branching ratios for oxidation of NH and HNO forming NO are decreased with less oxygen. As discussed above, those two pathways are the main routes for NO formation. Therefore, less NO is formed under fuel rich conditions. All of the above pathways are H, O, or OH relative reactions. Figure 8 presents the change of H, O, and OH radicals with the equivalence ratio. O and OH radicals are largely produced through chain branching reaction H+ O₂ = OH+O. Oxygen decreases with equivalence ratio increasing, which leads to less radicals. On the contrary, the combination of NH_i radicals becomes more significant, which eventually results in less NO formation and more consumption.

Figure 9 manifests the change of H, O, and OH radicals with syngas ratio. The oxidation of CO and H₂ provides more active radicals like H, OH, and O through reactions CO+ OH= H+ CO₂, H₂ + OH= H+ H₂O, and H+ O₂ = OH+O. Those radicals accelerate the oxidation of ammonia, leading to higher laminar flame speeds as shown in Fig. 6. Figure 10 illustrates the reaction pathways at x_syn = 0.3, 0.5, and 0.7 to investigate the effect of syngas ratio on ammonia oxidation. It can be seen that with more syngas, the NH_i combination declines apparently as NH₂ and NH are consumed through reaction NH₂ + O= HNO+ H and NH+ O= NO+ H, respectively with more O atom. In conclusion, reactions with H and O possess a larger chain branching ratio with the increase of the mole fraction of syngas. As to NO formation, more NH₂ radicals react with O, producing HNO. Moreover, further oxidation of HNO leads to the dominant production of NO. Furthermore, directly formation from NH increases with the increase of the mole fraction of syngas. On the contrary, consumption of NO decreases with syngas ratio increasing.

Figure 11 shows the sensitivity analysis of ammonia to identify the reactions which serve as the key point in the oxidation of ammonia. Chain branching reaction H+ O₂ = O+ OH has a large sensitivity coefficient of 0.52 at φ = 0.9 and 0.67 at φ = 1.5, respectively. As shown in Fig. 11, it is obvious that small radical reactions such as CO+ OH= CO₂ + H, OH+ H₂ = H+ H₂O, and O+ H₂ = H+ OH have dominant promoting effects on laminar flame speeds, resulting in the generation of active H and OH radicals. Reactions like H+ O₂ = HO₂ and H+ OH= H₂O have strong negative effects on laminar flame speeds because they consume the reactive H and OH radicals and produce stable radicals. Under the fuel lean condition, those reactions have larger sensitivity coefficients with sufficient oxygen. The H abstraction reaction NH₃ + OH= NH₂ + H₂O has a negative effect on the laminar flame speed as it consumes the OH radical, producing H₂O. Laminar flame speeds are far sensitive to NH₂ relative reactions. The combination of NH_i radicals like NH₂ + NH= N₂H₂ + H and NH₂ + NH₂ = N₂H₃ + H, plays a promoting role in oxidation of ammonia. Those reactions consume NH_i radicals and produce more active H atom to accelerate oxidation of ammonia. In contrast, NH₂ + NH₂ = N₂H₂ + H₂ and NH₂ + NH₂ = N₂H₄ have a prohibiting effect on the laminar flame speed. Under the fuel rich condition, this reaction class exhibits a larger effect with sufficient fuel radicals.

At different syngas ratios as shown in Fig. 12, the sensitivity of small radical reactions like CO+ OH= CO₂ + H, OH+ H₂ = H+ H₂O, O+ H₂ = H+ OH, and H+ OH= H₂O increases with syngas ratio increasing. On the other hand, ammonia oxidation is less sensitive to NH₂ relative reactions NH₂ + NH= N₂H₂ + H, NH₂ + NH₂ = N₂H₃ + H, and N₂H₂ + H₂ = NH₂ + NH₂, with more syngas.

3.5 Kinetic analysis of NO formation

First, the Han model was validated against the low-pressure flame data of ammonia/oxygen flame in Bian et al. [33], as shown in Fig. S3 in ESM. The experimental results agree well with the modeling predictions. The concentrations of reactants (NH₃ and O₂) and products (NO and N₂O) are predicted quite accurately. Therefore, this model was used for kinetic analysis of NO formation.

Figure 13 is a prediction of NO concentration at different temperatures, pressures, and syngas ratios using the Han model. The pathway analysis suggests that NO formation is closely related to H, OH, O, NH₂, NH, and N radicals. With the increase of temperature and decrease of pressure, the temperature of the adiabatic flame increases, in turn, promoting radical formation as shown in Figs. 14 and 15. The mole fractions of H, OH, O, NH₂, NH, and N radicals of flames increase with the temperature and decrease with the pressure, which finally leads to more NO.

NO formation changes non-monotonically as syngas ratio varies. This reveals an increase in NO production with the mole fraction of syngas ratio of up to 70%, but decreases with the syngas ratio of higher than 70%. The pathway analysis in Fig. 7 shows that the formation of NO can be categorized into three routes: (1) NH₂–HNO–NO, (2) NH₂–NH–NO, and (3) NH₂–NH–N-NO. Most of the NO is formed from the reaction HNO+ H= NO+ H₂ while HNO+ OH= NO+ H₂O and NH+ O= NO+ H reactions are secondary formation pathways, and HNO mainly comes from the oxidation of NH₂ and NH radicals. Therefore, OH, H, O, NH₂, NH, and N radicals are needed for NO formation. The mole fractions of H, OH, O, NH₂, NH, and N radicals of flames at φ = 0.9, 298 K, and 1 atm, at different syngas ratios can be found in Fig. S4 in ESM. It can be seen that the maximum concentrations of H, OH, and O radicals in flames are increased with the increase of the mole fraction of syngas, especially the H atom, which promotes the reactivity of reactions, like HNO+ H= NO+ H₂, HNO+ OH= NO+ H₂O, and NH+ O= NO+ H. On the other hand, NH₂, NH, and N radicals change non-monotonically as syngas ratio varies. They increase first and then decrease with syngas ratio of up to 50% owing to the decreasing content of NH₃ in the blend, which decreases the formation of NO. Therefore, NO formation peaks at around 70% syngas.

It is worth mentioning that with 90% syngas, NO formation exhibits a different trend from others. NO formation almost remains the same with the equivalence ratio larger than 1.2. To figure out the deep kinetic insight of NO formation, the mole fraction of H, O, OH, NH₂, NH, and N radicals are investigated as shown in Fig. S5 in ESM. The H and NH₂ radicals increase with more fuel while NH, N, O and OH radicals decrease with less oxygen. Therefore, the synthesis results in the trend of NO.

To identify the reactions that serve as bottlenecks in the formation of NO, a local sensitivity analysis is performed with the Han model for φ = 0.9 and φ = 1.5 in Fig. 16. The positive and negative sensitivity coefficients indicate the inhibition and promotion effect on NO formation, respectively. It is obvious that reactions H+ O₂ = O+ OH, CO+ OH= H+ CO₂, and H₂ + OH= H+ H₂O play important roles on NO formation. These reactions provide active radicals like H, OH, and O which accelerate NO formation. It can also be seen that ammonia related reactions NNH+ OH= NH₂ + NO and NH₂ + NH= N₂H₂ + H have a great promoting effect on NO formation. On the other hand, reactions H+ O₂ = HO₂ and H+ OH= H₂O have strong negative effects on NO formation as well as OH abstraction reactions. These reactions consume active radicals and prohibit NO formation. The NH_i combination exhibits a negative effect on NO formation as it consumes the NH_i radical and produces relative stable species which are not easily oxidized to produce NO. Small radical reactions have larger sensitivity coefficients under the fuel lean condition while ammonia related reactions possess larger ones under the fuel rich condition.

Figure 17 shows the effect of syngas ratio on sensitive reactions. The CO and H₂ related reactions, CO+ OH= CO₂ + H, OH+ H₂ = H+ H₂O, O+ H₂ = H+ OH, and H+ OH= H₂O shows larger sensitivity coefficients with the syngas ratio increasing. On the other hand, NO formation is more sensitive to NH₃ relative reactions NH₂ + NH= N₂H₂ + H, NNH+ OH= NH₂ + NO, and NH₃ + OH= NH₂ + H₂O, with more syngas.

4 Conclusions

The laminar flame speeds of ammonia mixed with syngas at high pressure and temperature and different syngas ratios were measured in a constant volume combustion bomb. The data obtained are fitted at different pressures, temperatures, syngas ratios, and equivalence ratios. The obtained equation can give an accurate prediction of laminar flame speeds under current conditions. Three kinetic models (Shrestha model, Glarborg model and Han model) were compared and validated against the current experimental data. The Han model has the best performance followed by the Shrestha and the Glarborg model.

Pathway, sensitivity, and radical pool analysis are conducted on NH₃ oxidation to find out deep kinetic insight. Ammonia is largely consumed by H-abstraction reactions with the O/H radicals then produces NH₂. The subsequent reactions of NH₂ can be categorized into three routes, i.e., NH₂–NH–N–N₂, NH₂–HNO–NO–N₂, and NH₂/NH–(N₂H₃–)N₂H₂–NNH-N₂. Similar to hydrocarbons, ammonia oxidation is sensitive to small radical reactions and fuel related reactions which produce and consume H, OH, and O radicals. Fuel-NO formation poses a great challenge to application of ammonia as fuel. NO production increases with syngas ratio of up to 70% and decreases at a syngas ratio higher than 70%. It is also worth mentioning that the main consumption channels and sensitive reactions of NO. NO formation can also be categorized into three routes, i.e., NH₂–HNO–NO, NH₂–NH–NO, and NH₂–NH–N-NO which are closely related to H, O, and OH radicals. Similar to ammonia oxidation, NO formation is sensitive to small radical reactions and ammonia related reactions.

References

[1] Valera-Medina A, Morris S, Runyon J, Ammonia, methane and hydrogen for gas turbines. Energy Procedia, 2015, 75: 118–123

[2] Balestri M, Cecchini D, Cinti V. Unconventional fuels experimental campaigns in gas turbine combustor at ENEL Sesta facility. In: Proceedings of ASME Turbo Expo 2004: Power for Land, Sea, and Air, Vienna, Austria, 2008: 121–128

[3] Kurata O, Iki N, Matsunuma T, Performances and emission characteristics of NH₃-air and NH₃CH₄-air combustion gas-turbine power generations. Proceedings of the Combustion Institute, 2017, 36(3): 3351–3359

[4] Iki N, Kurata O, Matsunuma T, Micro gas turbine firing kerosene and ammonia. In: Proceedings of ASME Turbo Expo 2015: Turbine Technical Conference and Exposition, Montreal, Canada, 2015

[5] Mei B, Zhang X, Ma S, Experimental and kinetic modeling investigation on the laminar flame propagation of ammonia under oxygen enrichment and elevated pressure conditions. Combustion and Flame, 2019, 210: 236–246

[6] Xiao H, Valera-Medina A, Bowen P J. Study on premixed combustion characteristics of co-firing ammonia/methane fuels. Energy, 2017, 140: 125–135

[7] Xiao H, Howard M, Valera-Medina A, Study on reduced chemical mechanisms of ammonia/methane combustion under gas turbine conditions. Energy & Fuels, 2016, 30(10): 8701–8710

[8] Wiseman S, Rieth M, Gruber A, A comparison of the blow-out behavior of turbulent premixed ammonia/hydrogen/nitrogen-air and methane-air flames. Proceedings of the Combustion Institute, 2021, 38(2): 2869–2876

[9] Issayev G, Giri B R, Elbaz A M, Combustion behavior of ammonia blended with diethyl ether. Proceedings of the Combustion Institute, 2021, 38(1): 499–506

[10] Zhao Y, Xie Y, Wang X, Energy balance analysis, combustion characteristics, and particulate number concentration-NO_x trade-off of a heavy-duty diesel engine fueled with various PODEn/diesel blends. Energy Conversion and Management, 2020, 225: 113489

[11] Whitty K J, Zhang H R, Eddings E G. Emissions from syngas combustion. Combustion Science and Technology, 2008, 180(6): 1117–1136

[12] Rota R, Antos D, Zanoelo E F, Experimental study and kinetic modelling of nitric oxide reduction with ammonia. Combustion Science and Technology, 2001, 163(1): 25–47

[13] Hayakawa A, Goto T, Mimoto R, Laminar burning velocity and Markstein length of ammonia/air premixed flames at various pressures. Fuel, 2015, 159: 98–106

[14] Han X, Wang Z, He Y, Experimental and kinetic modeling study of laminar burning velocities of NH₃/syngas/air premixed flames. Combustion and Flame, 2020, 213: 1–13

[15] Wang S, Wang Z, Elbaz A M, Experimental study and kinetic analysis of the laminar burning velocity of NH₃/syngas/air, NH₃/CO/air and NH₃/H₂/air premixed flames at elevated pressures. Combustion and Flame, 2020, 221: 270–287

[16] Mei B, Ma S, Zhang Y, Exploration on laminar flame propagation of ammonia and syngas mixtures up to 10 atm. Combustion and Flame, 2020, 220: 368–377

[17] The Combustion Research Group at UC San Diego. Chemical-kinetic mechanisms for combustion applications. University of California at San Diego, 2014, available at the website of

[18] Okafor E C, Naito Y, Colson S, Experimental and numerical study of the laminar burning velocity of CH₄-NH₃-air premixed flames. Combustion and Flame, 2018, 187: 185–198

[19] Zhang Y, Mathieu O, Petersen E L, Assessing the predictions of a NO_x kinetic mechanism on recent hydrogen and syngas experimental data. Combustion and Flame, 2017, 182: 122–141

[20] Shrestha K P, Seidel L, Zeuch T, Detailed kinetic mechanism for the oxidation of ammonia including the formation and reduction of nitrogen oxides. Energy & Fuels, 2018, 32(10): 10202–10217

[21] Han X, Wang Z, Costa M, Experimental and kinetic modeling study of laminar burning velocities of NH₃/air, NH₃/H₂/air, NH₃/CO/air and NH₃/CH₄/air premixed flames. Combustion and Flame, 2019, 206: 214–226

[22] Glarborg P, Miller J A, Ruscic B, Modeling nitrogen chemistry in combustion. Progress in Energy and Combustion Science, 2018, 67: 31–68

[23] Shrestha K P, Lhuillier C, Barbosa A A, An experimental and modeling study of ammonia with enriched oxygen content and ammonia/hydrogen laminar flame speed at elevated pressure and temperature. Proceedings of the Combustion Institute, 2021, 38(2): 2163–2174

[24] Yin G, Hu E, Huang S, Experimental and kinetic study of diisobutylene isomers in laminar flames. Energy, 2019, 170: 537–545

[25] Yin G, Gao Q, Hu E, Experimental and kinetic study on laminar flame speeds of formic acid. Combustion and Flame, 2020, 220: 73–81

[26] Hu E, Yin G, Ku J, Experimental and kinetic study of 2, 4, 4-trimethyl-1-pentene and iso-octane in laminar flames. Proceedings of the Combustion Institute, 2019, 37(2): 1709–1716

[27] Kelley A P, Liu W, Xin Y X, Laminar flame speeds, non-premixed stagnation ignition, and reduced mechanisms in the oxidation of iso-octane. Proceedings of the Combustion Institute, 2011, 33(1): 501–508

[28] Kee R J, Grcar J F, Smooke M D, PREMIX: a Fortran program for modeling steady laminar one-dimensional premixed flames. Sandia National Laboratories Report, SAND85–8249, 1985

[29] Yin G, Xu J, Hu E, Experimental and kinetic study on the low temperature oxidation and pyrolysis of formic acid in a jet-stirred reactor. Combustion and Flame, 2021, 223: 77–87

[30] Manton J, Elbe G V, Lewis B. Burning-velocity measurements in a spherical vessel with central ignition. In: Symposium (International) on Combustion, 1953, 4(1): 358–363

[31] Markstein G H. Nonsteady Flame Propagation. Oxford: Pergamon Press, 1964

[32] Varea E, Modica V, Renou B, Pressure effects on laminar burning velocities and Markstein lengths for Isooctane-Ethanol-Air mixtures. Proceedings of the Combustion Institute, 2013, 34(1): 735–744

[33] Bian J, Vandooren J, Tiggelen P J V. Experimental study of the structure of an ammonia-oxygen flame. Symposium (International) on Combustion, 1988, 21(1): 953–963

RIGHTS & PERMISSIONS

Higher Education Press