(593e) Chemical Kinetic Modelling of Ammonia-Hydrogen-Air Premixed Flames

Ammonia (NH₃) has been proposed as an alternative fuel for
a range of applications, such as internal combustion engines, gas turbines and
space propulsion systems [1,2]. Its production is cost-effective and can be achieved
through a number of methods, including from biomass and waste [3]. As a
flammable substance that does not contain carbon in its composition, its
combustion does not emit carbon dioxide (CO₂), which is the main
driver of global warming. Ammonia is also considered a safe hydrogen (H₂)
carrier due to its low reactivity and can be decomposed onsite into H₂
through catalytic processes [4]. However, NH₃ presents very low
flame speeds [5–7], being rather problematic to ignite and burn stably. Additionally,
NH₃ produces very high amounts of nitrogen oxides (NO_x), especially
under lean conditions, which are involved in atmospheric pollution, causing
respiratory problems and acid rain. One of the solutions proposed to enable its
use is co-firing ammonia with H₂ in order to improve flammability, and
thereby generate more stable flames with higher flame speed [8,9].

To develop systems suitable for using NH₃ and its blends with
H₂ in a clean and efficient way, it is necessary to understand its
combustion chemistry, which can be achieved through fundamental chemical
kinetics studies. The present work uses recently developed kinetic mechanisms [7,10,11]
in order to examine critical features associated with the combustion of NH₃
and its blends with H₂ such as ignition delay times and laminar flame
speeds. The simulations are compared with available experimental data. Ignition
delay time simulations were carried out for NH₃/O₂ mixtures
under zero-dimensional shock tube conditions, with 98-99% dilution in argon, under
pressure of 1.4, 11 and 30 atm, and temperature ranging from 1564 K to 2489 K. Flame
speed and NO_x formation were computed for ammonia/hydrogen/air
flames, with ammonia/hydrogen proportion ranging from 0 to 100%, under one-dimensional
freely propagating flame conditions, with premixed reactants injected at 298 K,
pressures of 1, 3 and 5 bar, and fuel/air equivalence ratios ranging from 0.5
to 1.5. All simulations were done using the Cantera package [12].

The mechanisms of Mathieu and Petersen [7], Otomo et al. [10] and
Okafor et al. [11] contain 34, 32 and 25 chemical species, and 159, 213, and
100 reactions, respectively. These mechanisms were reduced to contain 22, 24
and 21 species and 66, 72 and 51 reactions, respectively, in order to reduce
the computational demand. The reduced mechanisms based on Mathieu and Petersen
[7], Otomo et al. [10] and Okafor et al. [11] are referred to as the MP-R66,
the Ot-R72 and the Ok-R51 mechanisms, respectively. Figure 1 shows the
predicted ignition delay time of NH₃/O₂/Ar mixtures under
pressure of 11 atm and a range of temperature. The mechanisms of Otomo et al.
[10] and Mathieu and Petersen [7] show better agreement with experimental data
for the ignition delay times, which range from 20 µs at 2500 K to 2500 µs at 1500 K. The mechanism of Okafor [11] predicts higher values,
reaching 100 µs at 2500 K and
5000 µs at 1500 K. The reduced
mechanisms present slightly higher values than the original ones, reaching 30 µs at 2500 K and 4000 µs at 1500 K for Ot-72.

Figure 1. Shock tube ignition delay times for NH₃/O₂
mixtures with 99% dilution in argon at 11 atm for stoichiometric conditions.

The mechanism of Mathieu and Petersen [7] predicts the ignition
delay time of NH₃/O₂/Ar mixtures in better agreement with
the experiments than the other mechanisms do; however, this mechanism predicts
generally lower flame speeds in premixed NH₃/air flames at 1 bar and
298 K, Figure 2. On the contrary, the mechanism of Okafor et al. [11] and its
reduced mechanism Ok-R51 over-predict the ignition delay time and the flame
speeds, Figures 1 and 2. The mechanism of Otomo et al. [10] and its reduced one
Ot-R72 predict flame speeds and ignition delay time in between those from the other
mechanisms. In general, the reduced mechanisms predicts similar results as the
original ones do, except that the Ot-72 mechanism predicts a lower flame speeds
than those predicted by the original mechanism in the lean and rich mixtures,
Figure 2. It is worth mentioning that all mechanisms predict peak laminar flame
speed at slightly rich mixtures, around equivalence ratio of 1.1. This is
comparable with the flame methane/air flames that have a peak flame speed at
equivalence ratio of 1.06.

Figure 2. Laminar flame speed of premixed NH₃/air flames
at 1 bar and 298 K.

For NH₃ and H₂ blends (Fig. 3) at
stoichiometric condition all mechanisms predict lower flame speeds than the
measured ones for lower H₂ content. The mechanism of Okafor et al.
[11] predicts higher flame speeds for low H₂ content, while the mechanism
of Mathieu and Petersen [7] shows the poorest agreement with the experiments
for mixtures with low H₂ content. The mechanism of Otomo et al. [10]
predicts flame speeds in between those predicted by the other mechanisms. The reduced
mechanisms replicate the results of their respective original ones fairly well,
Fig.3. It should be pointed out that the laminar flame speed of stoichiometric NH₃/H₂/air
flames increase approximately exponentially with the mole fraction of hydrogen
in the fuel blends. Thus, it is an effective approach to improve the flame
speeds of ammonia flames by blending hydrogen in the fuel mixtures.

Figure 3. Laminar flame speed for stoichiometric premixed NH₃/H₂/air
flames at 1 bar and 298 K, as a function of the mole fraction of hydrogen in
the ammonia/hydrogen fuel blends.

All mechanisms predict very high NOx emissions for pure ammonia/air
combustion (Fig. 4). The reduced mechanisms present similar behaviour, except
for Ot-72. The predicted peak NOx is in lean mixtures, with equivalence ratio
about 0.9. The predicted peak NOx mole fraction varies from 2800 ppm to 4600
ppm, indicating a huge scatter of predictions from different mechanisms. For
ammonia-hydrogen blends (Fig. 5) the predicted NOx increases with the content
of hydrogen in the mixture, owing to the elevated temperature of the mixture, until
the mole fraction of hydrogen reaches 80% in the ammonia/hydrogen blends.
Further increase in the hydrogen content results in a decrease in NOx
production, owing to the decreased content of ammonia in the blends, Fig. 5.
The reduced mechanisms predict similar NOx results as those from their original
mechanisms, except for Ok-51 at conditions of high hydrogen content in the
blends. It should be pointed out that no experimental data available in the
literature for NOx in ammonia flames.

In summary, the results shown above clearly indicate that the neat
ammonia flames have rather low flame speed, and blending hydrogen with ammonia
is an effective approach for improving the flame speed of ammonia flames.
However, NOx emission is significantly higher in ammonia/hydrogen flames. The
present study shows that the presently available chemical kinetic mechanisms in
the literature predicts rather scattered ignition delay time, laminar flame
speed, and NOx emissions in ammonia flames.

Figure 4. NO_x concentrations for premixed NH₃/air
flames at 1 bar and 298 K.

Figure 5. NO_x concentrations for stoichiometric premixed
NH₃ + H₂/air flames at 1 bar and 298 K.

References

[1]               
Zamfirescu C, Dincer I. Ammonia as a green fuel
and hydrogen source for vehicular applications. Fuel Process Technol
2009;90:729–37.

[2]               
Valera-Medina A, Morris S, Runyon J, Pugh DG,
Marsh R, Beasley P, et al. Ammonia, Methane and Hydrogen for Gas Turbines. Energy
Procedia 2015;75:118–23.

[3]               
Bicer Y, Dincer I, Zamfirescu C, Vezina G, Raso
F. Comparative life cycle assessment of various ammonia production methods. J
Clean Prod 2016;135:1379–95.

[4]               
Mukherjee S, Devaguptapu S V., Sviripa A, Lund
CRF, Wu G. Low-temperature ammonia decomposition catalysts for hydrogen
generation. Appl Catal B Environ 2018;226:162–81.

[5]               
Hayakawa A, Goto T, Mimoto R, Arakawa Y, Kudo T,
Kobayashi H. Laminar burning velocity and Markstein length of ammonia/air
premixed flames at various pressures. Fuel 2015;159:98–106.

[6]               
Brackmann C, Alekseev VA, Zhou B, Nordström E,
Bengtsson PE, Li Z, et al. Structure of premixed
ammonia + air flames at atmospheric pressure: Laser diagnostics and kinetic
modeling. Combust Flame 2016;163:370–81.

[7]               
Mathieu O, Petersen EL. Experimental and
modeling study on the high-temperature oxidation of Ammonia and related NOx
chemistry. Combust Flame 2015;162:554–70.

[8]               
Ichikawa A, Hayakawa A, Kitagawa Y, Kunkuma
Amila Somarathne KD, Kudo T, Kobayashi H. Laminar burning velocity and
Markstein length of ammonia/hydrogen/air premixed flames at elevated pressures.
Int J Hydrogen Energy 2015;40:9570–8.

[9]               
Li J, Huang H, Kobayashi N, Wang C, Yuan H.
Numerical study on laminar burning velocity and ignition delay time of ammonia
flame with hydrogen addition. Energy 2017;126:796–809.

[10]           
Otomo J, Koshi M, Mitsumori T, Iwasaki H, Yamada
K. Chemical kinetic modeling of ammonia oxidation with improved reaction
mechanism for ammonia/air and ammonia/hydrogen/air combustion. Int J Hydrogen
Energy 2018:1–11.

[11]           
Okafor EC, Naito Y, Colson S, Ichikawa A, Kudo
T, Hayakawa A, et al. Experimental and numerical study of the laminar burning
velocity of CH4–NH3–air premixed flames. Combust Flame 2018;187:185–98.

[12]           
Goodwin DG, Moffat HK, Speth RL. Cantera: An
Object-oriented Software Toolkit for Chemical Kinetics, Thermodynamics, and
Transport Processes 2017.