Applied Thermal Engineering 259 (2025) 124748

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Numerical and experimental investigation of the correlation between inlet

flow rates, temperature, and NOx emissions in pure hydrogen combustion
Dang Khoi Le, Hyunguk Kwon * , Min Jung Lee *
Department of Future Energy Convergence, Seoul National University of Science and Technology, 232 Gongneung-ro, Nowon-gu, Seoul 01811, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Keywords: Many industrial burners currently in use have been designed for fossil fuels and have not been tested with
Hydrogen combustion hydrogen. The successful utilization of hydrogen fuel in industrial applications demands the optimization of
NOx formation operating conditions. Optimizing the inlet flow rate is important particularly for industrial burners to ensure safe
Inlet flow rate
and efficient operation over a long period. This study investigates the effects of inlet flow rate on temperature
Industrial burners
distribution and NOx emissions in a multi-purpose pure hydrogen-fueled industrial burner through numerical
Carbon-free
simulations and experiments. The numerical simulations, employing the realizable k − ε turbulence model and
the eddy dissipation concept for turbulence and combustion modeling, are validated against experimental data,
achieving a mean absolute percentage error of approximately 5 % for temperature and less than 2 % for NOx
emissions at 4 % O2 . The results demonstrated that increasing the inlet flow rate enhances reaction intensity and
heat release, subsequently increasing combustion temperatures and promoting NOx formation. However, with a
substantial increase in the inlet flow rate, the rate of temperature growth diminishes, and NOx emissions reach a
plateau. This suggests that the burner is approaching maximum capacity, with the combustion process becoming
limited by the burner’s ability to efficiently mix and react the fuel and oxidizer. Finally, an optimal inlet flow rate
of 50.4 LPM for hydrogen in the fuel inlet and 200 LPM for air in the oxidizer inlet, balancing temperature and
NOx emissions, was identified through the weighted sum performance indicator proposed in this work. This study
provides critical insights into pure hydrogen burner performance, advancing previous efforts by offering quan-
titative data essential for improving burner efficiency and reducing harmful emissions in carbon-free
combustion.

1. Introduction However, co-firing still generates carbon emissions and thus cannot
serve as the ultimate solution. A shift towards pure hydrogen combus-
Alternative fuels with minimal environmental effects are being tion is imperative for achieving a decarbonized society. Conversely,
developed as a result of growing concerns about climate change and burning pure hydrogen has limitations in terms of safety risks and
sustainability challenges with the current energy system. Hydrogen harmful pollutants. Compared to traditional fossil fuels, the high
stands out as a key factor in shaping the future of energy [1,2], as its burning velocity and wide flammability range of hydrogen increase the
ability to eliminate direct carbon emissions during combustion. It is risk of flashback and accidental explosions. In addition, the combustion
feasible to generate hydrogen as an energy carrier using renewable of pure hydrogen typically generates high-temperature flames, resulting
sources, such as wind and solar energy [3–5]. The large-scale CO2-free in the production of large amounts of thermal nitrogen oxides (NOx )
hydrogen production has been improving due to the increasing maturity through the oxidation of nitrogen in the air.
of electrolysis technology. Co-firing hydrogen with light hydrocarbons, Recently, reducing NOx emissions in hydrogen combustion has
such as natural gas, has been widely studied [6–9]. This approach pro- become a growing challenge for the combustion community. Optimizing
duces fewer carbon emissions compared to using pure hydrocarbon burner geometry is a possible solution for NOx reduction. Burners
fuels, leverages existing infrastructure. It establishes hydrogen-fueled incorporating complex flow regimes have been proposed, such as flame-
burners/engines as an appealing prospect for near-future, multi-scale sheet burners [11], micro-mixing burners [12], swirl burners [13], and
solutions targeting the decarbonization of several industries [10]. more recently, partially premixed bluff body burners [14,15]. These

* Corresponding authors.
E-mail addresses: hkwon@seoultech.ac.kr (H. Kwon), mj.lee@seoultech.ac.kr (M.J. Lee).

https://doi.org/10.1016/j.applthermaleng.2024.124748
Received 6 August 2024; Received in revised form 17 October 2024; Accepted 28 October 2024
Available online 30 October 2024
1359-4311/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

burners are characterized by highly optimized flow patterns and com- the operating conditions of the baseline case in a multi-purpose burner
plex designs, aiming to achieve optimal operational performance with through experiments and numerical simulations. The experimentally
minimal NOx emissions and enhanced efficiency. Cho et al. [16] also measured temperature and NOx concentration are used to select the
showed that the mixing uniformity of fuel and air inside the burner is a appropriate numerical models and validate the simulation results. Next,
key factor in controlling NOx production. They modified the burner the numerical models are used to analyze the temperature distribution
geometry to increase the time for mixing of fuel and air by changing the and NOx emissions associated with varying inlet flow rates of the burner.
interval between fuel holes, which resulted in NOx reduction. Modifying An optimal inlet flow rate is determined by balancing the objectives of
an industrial burner geometry involves substantial trial-and-error, and maximizing temperature and minimizing NOx emissions using a pro-
may require the development of new infrastructure, which entails sig- posed weighted sum performance indicator. This study provides a
nificant cost and time investment. deeper understanding of the effects of increasing inlet flow rates on pure
It is therefore necessary to consider and test the use of industrial hydrogen-fueled burners, establishing a foundational database for
burners, which have been used with fossil fuels, for hydrogen combus- optimizing burner operating conditions in future work.
tion without any burner geometry modifications. Since the transition
from fossil fuels to hydrogen fuel can lead to very different combustion 2. Methods
characteristics, such as flame stability, combustion efficiency, and NOx
formation, optimizing the operating condition is critically required. 2.1. Experimental apparatus
There is general agreement that lean-burn operation of hydrogen com-
bustion, by varying the equivalence ratio, can achieve a reduction in 2.1.1. Burner design
temperature and consequently thermal NOx formation [17–21]. On the In this study, a multi-purpose industrial burner was used to investi-
other hand, research on the relationship between inlet flow rates, while gate pure hydrogen combustion through experimental measurements
maintaining a constant fuel–air equivalence ratio, and the distribution of and numerical simulations. Fig. 1 displays the schematic of burner’s
temperature and NOx emissions in pure hydrogen-fueled industrial design, dimensions, and operational principles. Initially, the supplied
burners is still very limited. Identifying the optimal inlet flow rate is hydrogen and air flow separately, with hydrogen delivered through the
important particularly for large-sized industrial burners. The operation centrally positioned fuel inlet tube and air through the oxidizer inlet
of industrial burners requires careful and proper preparation for start- tube. Hydrogen and air then flow into the combustion zone through
up. Therefore, ensuring safe and efficient operation over a long with radially arranged holes, undergoing three stages of mixing before igni-
the proper inlet flow rate is crucial. tion and combustion. Subsequently, the combustion and exhaust gases
The aim of this study is to determine the relation of inlet flow rate, pass through the burner exit, where they are either utilized for specific
one of the most important parameters of burner operating conditions, to applications to provide heat or subjected to research analysis.
temperature distribution and NOx emissions of a pure hydrogen-fueled,
multi-purpose industrial burner. This industrial burner has been exten- 2.1.2. Experimental procedure
sively utilized with fossil fuels, such as natural gas; though, its appli- Fig. 2 illustrates the schematic diagram of the hydrogen burner test
cability to hydrogen fuel has not been investigated. We first determine rig. A pure hydrogen flow rate of 25.2 LPM (liter per minute) was

Fig. 1. Schematic of design, dimensions, and operational principles of the multi-purpose pure hydrogen-fueled burner.

2
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

Fig. 2. Schematic diagram of the hydrogen burner test rig.

injected at the fuel inlet. Note that the unit LPM is used in this work to directed towards the TESTO 350 K Gas analyzer for the measurement of
represent the volumetric flow rate under standard conditions (0 ◦ C and NOx emissions.
1 atm), which is equivalent to SLPM (standard liter per minute). To
ensure safety and prevent flashback, a flashback arrestor was inten-
tionally placed in the hydrogen inlet pipe. The primary air flow, at a rate 2.2. Numerical simulations
of 100 LPM, was injected into the oxidizer inlet for combustion. The
secondary air at a rate of 50 LPM was introduced to cool down the 2.2.1. Governing equations, turbulence, and combustion models
temperature at the outlet. The secondary inlet can also be used for The flow of reacting gas is governed by the Navier-Stokes equations,
multiple purposes, including the complete combustion of unburned which include continuity, momentum, energy, and species transport
fuels. All inlet flows were regulated by mass flow controllers (MFCs) and equations.
monitored through an RS232 Readout box, which were controlled by
LabVIEW software. • Continuity (conservation of mass) equation
The primary air and hydrogen flowed into the burner, where ignition ∂ρ
initiated primary combustion. As the combustion progressed, a flame is v) = 0
+ ∇ • (ρ→ (1)
∂t
formed within the burner and then exhausted into the flame zone, acting
as the secondary combustion zone. The flame temperatures were where ρ is the density, →
v is the velocity vector, and t is time.
measured using a GL240 Data logger at five distinct points within the
flame zone, labeled P01 through P05, positioned at 10 cm intervals from • Conservation of momentum equation
the top of the cover box (the black outer box). The thermocouples used ∂ →
for measuring flame temperature have been calibrated prior to the ex- (ρ v ) + ∇ • (ρ→
v→v ) = − ∇P + ∇ • (τ) (2)
∂t
periments to ensure accurate temperature readings. The calibration was
carried out by a domestic professional institution, and the measurement where P is the static pressure and τ is the stress tensor. Because of the
uncertainty of all five thermocouples was within 1.5 ◦ C. There is also an high velocity of the inlet jet, the reacting flows fall inside a regime
approximately estimated 5 cm gap between the burner exit and the top dominated by momentum, where the influence of body forces such as
of the cover box, making the location of point P01 about 15 cm from the gravity and external forces in the momentum equation is neglected.
burner exit. The flue gas exited through the vent, with a portion of it
• Energy equation

[ ( )] [ ( )] ( )
∂ v2 v2 ∑ →
ρ e+ + ∇ • ρv h + = ∇ • keff ∇T − h J
j j + τ eff • v
→ + Sh (3)
∂t 2 2 j

3
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

the EDC can be found in [30]. Due to the stiffness of kinetic equations
→
where keff is the effective conductivity, Jj is the diffusion flux of and large number of species and reactions, detailed chemical kinetic
species j, and Sh is the heat generation rate from chemical reactions. mechanisms require small time steps to avoid numerical instability,
which leads to the development of various reduced kinetic models
• Species transport equation [39–44]. GRI-Mech 3.0 [44] is widely recognized as a prominent kinetic
model widely employed in combustion research. Hong et al. [45] pro-
∂ →
v Yi ) = − ∇ • Ji + Ri
(ρYi ) + ∇ • (ρ→ (4) posed an updated H2 /O2 model based on recent shock tube measure-
∂t ments, enhancing rate constants for several reactions. Burke et al. [46]
updated the H2 /O2 kinetic model based on Li et al. [39], aiming to
where Yi is the mass fraction of species i and Ri is the net rate of pro-
address recent advancements in rate constant treatment and resolve
duction of species i by chemical reaction.
discrepancies in dilute, high-pressure flames. Numerous reduced
Reynolds-averaged Navier Stokes (RANS) models have been widely
chemical kinetic mechanisms have been evaluated to identify a model
used to model turbulent flow in combustion, its applicability in
that can offer accurate predictions while maintaining affordable
capturing unsteady combustion phenomena such as flashback, blow-off,
computational resources. The chemical kinetic mechanisms proposed by
lift-off, and vibration is limited [22]. Recently, Direct Numerical Simu-
Burke et al. [46] meet the aforementioned criteria and have been
lation (DNS) and Large Eddy Simulation (LES) have become popular due
selected to model hydrogen combustion in this study. In addition, dif-
to their capability to resolve the unsteady behavior of turbulent com-
ferential diffusion effects were accounted for by calculating species-
bustion, which requires high spatial and temporal resolution [23].
specific transport properties to ensures accurate simulation of
Conversely, they are computationally expensive. Considering the large
hydrogen combustion, as the non-unity Lewis number significantly in-
computational requirement for modeling chemical reactions, reducing
fluences mixing. These transport properties, derived from the Burke
computational costs has therefore been prioritized. It results in the se-
model, have been widely validated in hydrogen combustion simulations
lection of the Reynolds-Averaged Navier-Stokes (RANS) model for tur-
[47–49]. However, this kinetic model does not include the formation of
bulence modeling. The realizable k-epsilon (k − ε) turbulence model
nitrogen oxides (NOx ). In order to predict the NOx emissions, the Burke
[24], chosen for turbulence modeling in this study, has been validated
model was modified by incorporating NOx formation (referred to as
for its ability to capture flow behaviors in turbulent combustion by many
modified Burke model from now). This modification involves inte-
studies [12,18,25–29]. The realizable k − ε model is based on transport
grating additional reaction pathways and rate expressions for the for-
equations for the turbulence kinetic energy k (equation (5) and its
mation of NOx , based on the mechanism proposed by Zeldovich [50].
dissipation rate ε (equation (6) [30].
Detailed mechanisms, thermodynamic data, and transport data files are
[( ) ]
∂ ∂ ( ) ∂ μ ∂k provided in the Supplementary materials. All numerical simulations
(ρk) + ρkuj = μ+ t + Gk + Gb − ρε − YM (5)
∂t ∂xj ∂xj σk ∂xj were implemented using the CFD commercial software ANSYS Fluent
v23.2.
[( ) ]
∂ ∂ ( ) ∂ μt ∂ε ε2 ε
(ρε) + ρεuj = μ+ + ρC1 Sε − ρC2 √̅̅̅̅̅ + C1ε C3ε Gb
∂t ∂xj ∂xj σε ∂xj k + νε k 2.2.2. Boundary conditions and numerical schemes
(6) To accurately capture the complex flow dynamics and combustion
characteristics, three-dimensional (3D) simulations were performed.
The definitions and values of parameters and coefficients for the above Fig. 3 illustrates the geometry model and boundary conditions (BCs) of
equations can be found in reference [30], which are not repeated here. the numerical simulation. The fuel-to-air equivalence ratio (ϕ) is a
Along with turbulence, specialized treatment is necessary for the critical parameter in combustion. To maintain an appropriate temper-
chemical reactions in combustion, which are described by chemical ki- ature for minimizing NOx formation, ϕ was fixed at 0.6 in this work.
netic mechanisms. Several attempts have been undertaken to determine Using this ϕ value as a basis, a reference case of operating conditions was
the appropriate combustion modeling method and chemical kinetic selected, comprising a hydrogen volume flow rate of 25.2 LPM to the
model in this study. The Eddy Dissipation Concept (EDC) model [31,32] fuel inlet, an air volume flow rate of 100 LPM to the oxidizer inlet, and
is selected for combustion modeling due to its consistent simulation 50 LPM to the cooling inlet. This configuration was employed in both
results with experimental data found in the literature [18,29,33–38]. numerical simulation and experimental work to validate the numerical
The EDC extends the capabilities of the eddy-dissipation model by model. Table 1 summarizes the BCs and operating conditions of the
incorporating detailed chemical kinetic mechanisms into turbulent reference simulation case.
flows [31]. The net rate of production of species i (Ri ) is equation (4) can As mentioned in section 2.2.1, the realizable k − ε model and EDC
be computed by EDC as. were selected for modeling turbulence and combustion, respectively.
2
The Discrete Ordinates (DO) was used to model heat radiation
ρ ξ* ( * ) [18,36,51]. This model effectively handles both surface-to-surface and
Ri = ( ) Y i − Yi (7)
*3
τ 1− ξ
* participating radiation, including semi-transparent walls, across a wide
range of optical thicknesses. It offers moderate computational cost with
It is assumed that reactions occur within small turbulent structure, typical angular discretizations and low memory requirements, while
where ξ* is the length fraction of fine scales. also accommodating gray or non-gray radiation using a gray-band
model. To accurately model heat transfers at the walls, we employed
( )1
νε 4
(8) the Standard Wall Functions (as mentioned in Section 2.2.3). This model
ξ* = Cξ
k 2 utilizes a law-of-the-wall approach, combining a linear law for the
thermal conduction sublayer and a logarithmic law for the turbulent
where Cξ is the volume fraction constant equal to 2.137, and ν is the region. This comprehensive approach enhances the accuracy of thermal
kinematic viscosity. Species are assumed to react within the fine struc- performance predictions in combustion simulations by effectively
tures over a specific time scale. addressing both conduction and turbulence effects. Detailed principles
and equations for the DO radiation model, and Standard Wall Functions
(ν)1
τ* = Cτ
2 (9) can be found in reference [30], so they are not repeated here. “Coupled”
ε was chosen for the pressure–velocity coupling scheme due to its effec-
tiveness in handling the strong coupling between pressure and velocity
where Cτ is the time scale constant equal to 0.4082. More details about
fields, especially in turbulent and reactive flows, ensuring accurate

4
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

Fig. 3. The geometry model and boundary conditions of the numerical simulation.

in turbulent and reactive combustion simulations. Converging solutions

Table 1
in reacting flows is challenging due to the strong coupling between
Boundary conditions and operating conditions of the reference simulation case.
mass/momentum balances and species transport equations, especially in
Boundary Type Material Operating condition combustion where reactions cause significant heat release and flow ac-
Fuel Inlet Velocity inlet Hydrogen 5.925 m/s (velocity) celerations. These coupling issues are best addressed by using a two-step
Oxidizer Inlet Velocity inlet Air 23.513 m/s (velocity) solution procedure. The simulation procedure starts by solving the flow,
Cooling Inlet Velocity inlet Air 0.091 m/s (velocity)
energy, and species equations with reactions disabled, forming a “cold-
Outlet Pressure outlet Air 0 Pa (gauge pressure)
Walls No-slip wall Steel & glass 300 K (temperature)
flow” or unreacting flow pattern. Once the basic “cold-flow” is estab-
lished, reactions are enabled, following that an ignition source is
patched to initiate combustion. The simulation then proceeds, utilizing
predictions of flow behavior, flame dynamics, and combustion charac- the cold-flow solution as the initial condition for the combusting system
teristics [30]. The Second Order Upwind (SOU) scheme was selected for [30]. Although the objective is to achieve steady-state conditions, the
spatial discretization of momentum, turbulence, and species equations complex geometry necessitates the use of transient simulations to ensure
due to its ability to retain accuracy while minimizing numerical diffu- convergence. The time step for the transient simulation is set to 3E-04 s
sion, which is particularly beneficial for resolving complex flow features to ensure a balance between computational cost and accuracy. This time

Fig. 4. The hybrid grid of the numerical simulation.

5
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

step allows for residuals of 1E-04 for the continuity equation and 1E-05 experimental flame. The results demonstrate a close agreement between
for the remaining equations. The total simulation time is approximately the predicted flame structure and experimental observations, confirm-
2.5 s, determined by monitoring the area-weighted average of the outlet ing the predictive capability of the numerical model.
temperature. Once the temperature stabilizes and no longer fluctuates Fig. 6.d shows a comparison between the predicted temperatures and
over subsequent time steps, the flame is considered to have reached a the experimental measurements at five distinct points, labeled p01 to
steady-state. p05 (as illustrated in Fig. 2 and Fig. 6.c). Generally, the numerical
simulation matches well with experimental data, insignificantly over-
2.2.3. Grid generation and independency analysis estimating the temperature at P01 and slightly underestimating it from
Due to the complexity of the burner’s geometry, it was divided into P02 to P05. To quantitatively evaluate the difference between the nu-
multiple zones for grid generation. A hybrid grid was employed in the merical simulation and the experimental measurements, the mean ab-
numerical simulation, consisting of structural hexahedral mesh for solute percentage error (MAPE) and root mean square (RMS) values
simple geometry zones and tetrahedral mesh for complex geometry were calculated as follows.
zones, as shown in Fig. 4. In addition, an O-grid meshing technique was ⎧
utilized for circular geometry regions to enhance grid quality. The walls ⎪ ⃒ ⃒
⎪
⎪
⎪ 1∑5 ⃒⃒Texp sim ⃒
P0i − TP0i ⃒
significantly influence turbulent flows. Wall functions rely on the uni- ⎪
⎪ MAPE = exp ⃒ × 100%
⎪
⎨ 5 i=1 ⃒ TP0i
versal law of the wall, which hypothesizes that the velocity distribution (11)
near a wall is consistent across most turbulent flows. This study ⎪ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
⎪
⎪
⎪ 1∑5 ( exp )2
employed the standard wall function, which requires each wall-adjacent ⎪
⎪ RMS = T − Tsim
⎪
⎩ 5 i=1 P0i P0i
cell’s centroid is located within the log-law layer, 30 < y+ < 300. y+ is
the dimensionless wall distance given as [30].
yuτ Where Texp sim
P0i and TP0i are the experimental temperature and predicted
y+ = (10) temperature at point P0i, respectively. Table 3 summarizes the calcu-
ν
lated MAPE and RMS values. The MAPE is around 5 %, indicating a good
Where y is the absolute distance from the wall, uτ is the friction velocity, agreement between the numerical simulation and experimental mea-
ν is the kinematic viscosity. In this work, the near-wall mesh was surements. The low RMS values further support the accuracy of the
generated with an average dimensionless wall distance y+ ≅ 100, and numerical model in predicting temperature distributions.
maximum y+ less than 300. To evaluate the NOx emissions, the NOx measured on a dry basis by
Appropriate grid resolution significantly influences the accuracy of the TESTO 350 K Gas analyzer (Fig. 2) was converted to NOx at 4% O2 .
numerical simulations. Therefore, a grid independence analysis should This conversion is critical for standardizing the NOx measurements to a
be conducted to select the optimal grid resolution. Table 2 summarizes common reference point, thereby enabling meaningful comparisons
four grid resolution cases evaluated to determine the most appropriate with regulatory limits and findings from other studies. NOx emissions
grid size. are typically reported in Parts Per Million (PPM), and excess air in flue
Fig. 5 compares the temperature at the first experimental measure- gases can dilute these measurements, complicating their interpretation.
ment point P01 (see section 2.1) and the area-weighted average mass Regulatory authorities often require a standardized flue gas oxygen level
fraction of H2 O at the outlet across the four different grid resolution to ensure accurate reporting. In our research, the TESTO 350 K Gas
cases. Generally, both variables decrease as the number of grid cells analyzer is designed and calibrated specifically to measure NOx emis-
increases. The temperature at the first measurement point was experi- sions at 4% O2 , consistent with the national standards in South Korea. By
mentally measured as 975.5 ◦ C. This demonstrates that the predicted converting the NOx emissions to this standardized basis, we ensure that
values become more accurate with an increase in the number of grid our findings are both reliable and compliant with established regulatory
cells (finer grid). The predicted values display a minor difference be- frameworks, thus allowing for a precise assessment of the combustion
tween Grid 3 and Grid 4 (less than 1 %), indicating that the solution has system’s environmental performance. In the numerical simulation, the
come to be independent of the grid. Considering the computational cost dry NOx was calculated at the outlet and similarly converted to NOx at
and these findings, Grid 3 is supposed to be the most appropriate grid 4% O2 as follows.
resolution for numerical simulations in this study.
cmeasured
NOx (ppm) × (20.9% − 4%)
c4% O2
NOx (ppm) = (12)
3. Validation of the numerical models 20.9% − mmeasured
O2

Validating numerical models is essential for ensuring the accuracy where c4% O2
NOx (ppm) is the NOx at 4% O2 in parts per million (ppm),
and reliability of computational simulations. The experimental study of cmeasured
NOx (ppm) is the measured NOx in ppm, and mmeasured O2 is the
reference case (refer to Table 1) was used to validate the numerical measured mole fraction of O2 . Table 3 summarizes the predicted and
models. In order to visualize the numerical simulation results, a vertical experimental values of NOx at 4% O2 , along with their corresponding
cross-sectional plane through the center of the fuel and oxidizer inlets error. The error is less than 2 %, demonstrating the numerical model’s
(labeled V-plane), along with several horizontal cross-sectional planes capability to accurately predict NOx emissions.
(labeled Z1-plane to Z7-plane), were defined (refer to Supplementary In summary, the results confirm that the numerical model can
materials for visualizations of these planes). accurately predict temperature distribution and NOx emissions, proving
In combustion, flame structure is one of the most important char- it a good candidate for further investigations of the hydrogen burner at
acteristics, significantly impacting key aspects such as heat release, different inlet flow rates.
pollutant formation, and overall efficiency. Fig. 6 illustrates the nu-
merical prediction of temperature distribution along with the 4. Effects of inlet flow rate on temperature and NOx emissions

To investigate the effects of varying inlet flow rates while main-

Table 2 taining the fuel-to-air equivalence ratio ϕ at 0.6, the power settings of
Summary of grid independence analysis. the mass flow controllers, which regulate the electrical power supplied
Case Grid 1 Grid 2 Grid 3 Grid 4 to the pumps controlling the inlet flow rates of hydrogen and air, were
Number of cells 696,553 836,278 1,104,494 1,526,270
altered. Table 4 shows the different combinations of fuel (hydrogen) and

6
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

Fig. 5. Comparison of (a) temperature at the first experimental measurement point (refer to section 3) and (b) area-weighted average mass fraction of H2 O at outlet
between four grid resolution cases.

Fig. 6. Simulated temperature distribution contours at (a) horizontal cross-sectional planes and (b) V-plane. (c) Experimental flame. (d) Comparison of the predicted
and the experimental temperatures at measurement points P01 to P05.

study of how variations in inlet flow rates influence the combustion

Table 3
process.
Summary of average, MAPE, and RMS values for temperature distribution and
NOx at 4% O2 , and error between numerical simulation and experimental
measurements. 4.1. Effects on temperature distribution
Numerical Experimental
simulations measurements The temperature distribution contours for different inlet flow rates at
Temperature Average 744.693 775.490
horizontal cross-sectional planes and V-plane are illustrated on the same
MAPE (%) 5.041 scale, as shown in Fig. 7. Generally, as the inlet power (flow rate) in-
RMS (◦ C) 39.527 creases, there is a corresponding rise in temperature levels. In scenarios
NOx c4% O2
NOx (ppm)
189.197 192.847 with lower power inputs (1.5 kW, 3.0 kW, and 4.5 kW), shorter flame
emissions Error (%) 1.893 lengths are observed, contrasting with situations of higher power inputs
(6.0 kW, 7.5 kW, and 9.0 kW), where longer flames are evident. In these
higher power scenarios, the maximum temperatures are observed within
oxidizer (air) flow rates achieved by changing the power input to the
the secondary combustion zone.
inlet pumps. The cooling air flow rate was maintained unchanged, as in
Fig. 8 shows the temperature distribution at measurement points P01
the reference case. For convenience, each simulation case was labeled
to P05 and the area-weighted average temperature at horizontal cross-
according to the power values of the inlet pumps, enabling a systematic
sectional planes for different inlet flow rates. In the primary

7
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

Table 4
Summary of combinations of hydrogen and primary air flow rates according to power values of the inlet pumps.
Power (kW) 1.5 3.0 4.5 6.0 7.5 9.0

Fuel flow rate (LPM) 8.40 16.80 25.20 33.60 42.00 50.40
Fuel inlet velocity (m/s) 1.975 3.950 5.925 7.900 9.876 11.851
Reynolds number 1.73E+2 3.47E+2 5.20E+2 6.93E+2 8.67E+2 1.04E+3
Oxidizer flow rate (LPM) 33.33 66.67 100.00 133.33 166.67 200.00
Oxidizer inlet velocity (m/s) 7.837 15.676 23.513 31.350 39.190 47.026
Reynolds number 5.01E+3 1.00E+4 1.50E+4 2.00E+4 2.51E+4 3.01E+4

Fig. 7. Temperature distribution contours for various inlet flow rates at horizontal cross-sectional planes (top) and V-plane (bottom).

Fig. 8. Distribution of (a) temperature at measurement points P01 to P05 and (b) area-weighted average temperature at horizontal cross-sectional planes for various
inlet flow rates.

8
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

combustion zone (z < 0 m), higher area-weighted average temperatures combustion zones. Fig. 10 demonstrates an in-depth analysis of the
are observed with lower power inputs (1.5 kW, 3.0 kW, and 4.5 kW). fuel–air equivalence ratio for different inlet flow rates. In Fig. 10.a, the
Conversely, in the secondary combustion zone (z > 0 m), an increase in equivalence ratios are calculated for the primary and secondary regions
power leads to elevated temperatures observed both in the area- both before ignition (cold-flow) and after the burner reaches steady-
weighted average and at measurement points. The lower temperature state combustion. In the primary region, the equivalence ratio during
with higher power inputs within the primary zone may be due to cold-flow (without combustion) decreases from 0.62 to 0.48 as the inlet
incomplete combustion. Afterward, the unreacted fuel and air mixture power increases from 1.5 kW to 6.0 kW, and it remains at 0.48 when the
passes through the burner exit and enters the secondary combustion inlet power is further increased to 9.0 kW. During steady-state com-
zone (z > 0 m), where it undergoes combustion. This combustion pro- bustion, the equivalence ratio rises from approximately 0.27 to a peak of
cess in the secondary zone results in higher temperatures observed in 0.3 as the inlet power increases from 1.5 kW to 4.5 kW. Beyond this
cases with high inlet flow rates. These findings provide evidence of a input, the ratio stabilizes at 0.3. These results suggest that further in-
shift in the combustion process from the primary combustion zone to the creases in flow rate do not significantly affect the local fuel–air mixture
secondary combustion zone as the inlet flow rate increases. It is note- in the primary zone. This plateau indicates that combustion efficiency in
worthy that the temperature does not exhibit a linear relationship with the primary zone is constrained by the air and fuel mixing capacity of the
the inlet flow rate. As the inlet flow rate increases linearly, the rate of burner. During cold-flow, a similar trend is observed in the secondary
temperature growth diminishes. This trend indicates that the tempera- region. When increasing the inlet power from 1.5 kW to 4.5 kW, the
ture will likely reach a peak and remain relatively constant, even with equivalence ratio increases slightly from 0.02 to 0.04. However, it then
further increases in the inlet flow rate, signifying that the burner reaches rises sharply to a peak of 0.39, remaining constant as inlet power in-
its maximum capacity. creases from 6.0 kW to 9.0 kW. This behavior is attributed to the primary
The mass fraction of H2 distribution contours for different inlet flow region’s limited capacity between inlet powers of 4.5 kW and 6.0 kW.
rates at horizontal cross-sectional planes and the V-plane are illustrated Notably, the equivalence ratio in the secondary zone remains zero
on the same scale, as shown in Fig. 9. It shows that as the inlet flow rate during steady-state combustion for all flow rates, indicating complete
increases, the distribution of H2 mass fraction spreads further down- combustion in this region. This implies that all available fuel is fully
stream. In other words, at lower inlet flow rates, the primary combustion consumed in the secondary zone, highlighting the effectiveness of the
zone (z < 0 m) predominantly consumes the fuel (hydrogen). burner design in utilizing fuel after combustion products exit the pri-
Conversely, higher inlet flow rates result in increased presence of un- mary region. These findings underscore the relationship between inlet
burned fuel within the primary combustion zone, which then flows into flow rate and combustion zone activity. At lower inlet flow rates (1.5 kW
the secondary combustion zone (z > 0 m) for combustion. This remark to 4.5 kW), the primary combustion zone dominates, as evidenced by
strengthens the notion that higher inlet flow rates lead to a shift in lower equivalence ratios during steady-state combustion. Conversely, at
combustion from the primary combustion zone to the secondary com- higher flow rates (6.0 kW to 9.0 kW), unburned fuel is increasingly
bustion zone. carried into the secondary zone. As a result, the burner shifts from
Following the analysis of the H2 mass fraction distribution, the relying primarily on the primary zone at lower flow rates to utilizing
fuel–air equivalence ratio emerges as a crucial factor influencing the both zones at higher flow rates, with the secondary zone compensating
combustion characteristics across both the primary and secondary for any incomplete combustion. Fig. 10.b illustrates the distribution of

Fig. 9. Mass fraction of H2 contours for various inlet flow rates at horizontal cross-sectional planes (top) and V-plane (bottom).

9
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

Fig. 10. Distribution of (a) fuel–air equivalence ratio in primary and secondary regions, and (b) across horizontal cross-sectional planes during steady-state com-
bustion for various inlet flow rates.

the fuel–air equivalence ratio across various horizontal cross-sectional 4.2. Effects on NOx emissions
planes before ignition (cold-flow), further supporting the trends
observed in Fig. 10.a. Similar increases in the equivalence ratio of sec- The mass fraction of NO distribution contours for different inlet flow
ondary region with increasing inlet power are comprehended. These rates at horizontal cross-sectional planes and the V-plane are illustrated
observations support earlier findings that increasing inlet flow rates on the same scale, as shown in Fig. 11. As expected, increasing the inlet
does not linearly enhance temperature or combustion efficiency. flow rate leads to a rise in, which consequently increases the mass
Instead, combustion shifts from the primary to the secondary zone, fraction of NO. In cases of lower inlet power, NO mass fractions mainly
explaining the plateau in temperature rise and defining the operational found in the primary combustion zone. In contrast with higher inlet flow
limits of the burner. rate cases, NO mass fractions are mainly observed in the secondary
combustion zone, matching with the temperature trends. This trend
corresponds to the relationship between temperature and thermal NOx ,
where thermal NOx formation increases significantly with higher

Fig. 11. Mass fraction of NO contours for various inlet flow rates at horizontal cross-sectional planes (top) and V-plane (bottom).

10
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

temperatures. Notably, it is observed that higher NO emissions near the

where nT and n−N 1 are the normalized average temperature at measure-
burner exit can be attributed to the high temperatures present in this
ment points in secondary combustion zone and normalized inverse NOx
region (Fig. 7), along with the non-central placement of the oxidizer
at 4% O2 at outlet of the burner, respectively. These variables are
inlet. This design results in an asymmetric distribution of nitrogen (N2 )
calculated as follows.
and oxygen (O2 ), leading to higher concentrations of excess N2 and O2 in
⎧
the area adjacent to the oxidizer inlet. The increased availability of O2 , ⎪ T − Tmin
⎪
⎪ nT =
coupled with elevated temperatures, enhances the formation of NOx ⎪
⎪
⎨ Tmax − Tmin
through thermal mechanism. This behavior is mainly observed in cases / [ / 4% O2 ] (14)
of low inlet power, where the uneven distribution of excess N2 and O2
⎪
⎪
⎪ −1 1 c4% O2
NOx (ppm) − 1 cNOx (ppm) min
⎩ nN = [ / 4% O2
⎪
⎪ ] [ / ]
promotes higher NOx emissions. However, as the inlet flow rate in- 1 cNOx (ppm) max − 1 c4% O2
NOx (ppm) min
creases, resulting in higher velocity flow, the distribution of excess N2
and O2 becomes more uniform due to improved mixing. Consequently, where T is the average temperature at measurement points in each inlet
the higher NOx distribution near the oxidizer inlet diminishes in cases of flow power case, Tmin and Tmax are the minimum and maximum average
higher inlet power, leading to a more uniform distribution of NOx temperature at measurement points across all inlet power cases,
emissions in this region. respectively. The c4% O2
NOx (ppm) is the NOx at 4% O2 at outlet of the burner
Fig. 12 illustrates the variation in the area-weighted average mass [ ] [ ]
fraction of NO across horizontal cross-sectional planes and the concen- in each inlet flow power case, 1/c4% O2
NOx (ppm) , 1/c4% O2
NOx (ppm)
min max
tration of NOx at 4% O2 at the outlet for different inlet flow rates. In the are the minimum and maximum inverse NOx at 4% O2 at outlet of the
primary combustion zone (z < 0 m), the area-weighted average mass burner across all inlet power cases.
fraction of NO for low inlet power scenarios is greater than that for high The wT and wN are the relative weights of importance of the nT and
inlet power scenarios. This effect is caused by the higher temperatures n−N1 , respectively. In this study, the importance of maximizing temper-
observed under these conditions, as previously mentioned. In the sec- ature and minimizing NOx emissions are assumed to be equal to each
ondary combustion zone (z > 0 m), the area-weighted average NO mass other so that wT = wN = 1. A higher weighted sum performance in-
fraction rises with increasing inlet flow rates. Conversely, there is an dicates better performance of the burner. The proposed weighted sum
insignificant difference in this value between inlet power levels of 6.0 performance indicator PWSM effectively captures the balance between
kW, 7.5 kW, and 9.0 kW. This results in a substantial growth in NOx at temperature and NOx emissions by integrating them into a single metric.
4% O2 at outlet when the inlet power increases from 1.5 kW to 6.0 kW. By maximizing the temperature in the secondary combustion zone while
The NOx at 4% O2 at outlet plateaus and fluctuates around 235 ppm as simultaneously minimizing NOx emissions, the PWSM offers a compre-
the inlet power increases from 6.0 kW to 9.0 kW. This phenomenon hensive assessment of burner performance. This approach not only
occurs for the reason that at higher inlet flow rates, the burner is already highlights the importance of temperature for combustion efficiency but
operating at its peak efficiency, and additional fuel and air do not lead to also emphasizes the need to reduce harmful emissions, aligning with
proportional increases in temperature. environmental standards and sustainability goals. Table 5 summarizes
the average temperature, normalized average temperature, inverse NOx
4.3. Optimal inlet flow rate at 4% O2 , normalized inverse NOx at 4% O2 , and weighted sum perfor-
mance of all inlet powers. The inlet power of 3.0 kW delivers the worst
Higher temperatures can promote more complete combustion of the performance (PWSM = 0.8) while the 9.0 kW achieves the best perfor-
fuel, typically resulting in improved combustion efficiency. On the other mance (PWSM = 1.017). Therefore, the optimal inlet flow rate is the 9
hand, higher temperatures result in increased NOx emissions. Therefore, kW case, with 50.40 LPM of hydrogen for the fuel inlet and 200 LPM of
the burner’s performance is characterized by a trade-off between air for the oxidizer inlet.
maximizing temperatures in the secondary combustion zone and mini-
mizing NOx emissions. Based on the Weighted Sum Method [52], the 5. Conclusions
weighted sum performance indicator PWSM is proposed for determining
the optimal inlet flow rate, given as. Hydrogen arises as a crucial alternative fuel for carbon emission
PWSM = wT × nT + wN × n−N1 (13) mitigation in various applications. This study investigated the effects of
increasing inlet flow rate on temperature and NOx emissions in a multi-

Fig. 12. Distribution of (a) area-weighted average mass fraction of NO at horizontal cross-sectional planes and (b) NOx at 4% O2 at outlet for various inlet flow rates.

11
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

Table 5
Summary of average temperature, normalized average temperature, inverse NOx at 4% O2 , normalized inverse NOx at 4% O2 , and weighted sum performance of all
inlet powers.
Power (kW) 1.5 3.0 4.5 6.0 7.5 9.0

T 463.754 633.911 744.693 786.190 828.550 859.581

nT 0.000 0.430 0.710 0.815 0.922 1.000
1/c4% O2
NOx (ppm)
9.96E− 03 6.33E− 03 5.29E− 03 4.31E− 03 4.20E− 03 4.29E− 03
n−N1 1.000 0.370 0.189 0.020 0.000 0.017
PWSM 1.000 0.800 0.899 0.835 0.922 1.017

purpose pure hydrogen-fueled burner. The numerical simulations uti- Data availability
lized the realizable k − ε turbulence model for modeling turbulence flow
and the EDC model for modeling combustion, which incorporated the Data will be made available on request.
Burke H2/O2 kinetic model with Zeldovich’s NOx formation reactions.
Experimental work was performed to validate the numerical model. The References
results established a strong agreement between numerical simulations
and experimental measurements, with a MAPE around 5 % for tem- [1] M. Ball, M. Wietschel, The future of hydrogen – opportunities and challenges☆, Int.
J. Hydrogen Energy 34 (2009) 615–627, https://doi.org/10.1016/j.
perature distribution and an error of less than 2 % for NOx at 4% O2 at ijhydene.2008.11.014.
outlet, proving it as an appropriate candidate for further investigations [2] P. Moriarty, D. Honnery, Hydrogen’s role in an uncertain energy future, Int. J.
into the hydrogen burner at varying inlet flow rates. Hydrogen Energy 34 (2009) 31–39, https://doi.org/10.1016/j.
ijhydene.2008.10.060.
Increasing inlet flow rates generally correlated with higher temper- [3] G. Cipriani, V. Di Dio, F. Genduso, D. La Cascia, R. Liga, R. Miceli, G. Ricco
atures and NOx emissions, emphasizing the sensitivity of combustion Galluzzo, Perspective on hydrogen energy carrier and its automotive applications,
dynamics to flow rate variations. Notably, lower inlet flow rates corre- Int. J. Hydrogen Energy 39 (2014) 8482–8494, https://doi.org/10.1016/j.
ijhydene.2014.03.174.
sponded to higher area-weighted average temperatures in the primary [4] C. Acar, I. Dincer, Testing and performance evaluation of a hybrid
combustion zone, whereas higher inlet flow rates resulted in higher photoelectrochemical hydrogen production system, Int. J. Hydrogen Energy 42
area-weighted average temperatures in the secondary combustion zone. (2017) 3605–3613, https://doi.org/10.1016/j.ijhydene.2016.11.044.
[5] H.Y. Kim, N. Il Kim, Optimized global reaction mechanisms for H2, CO, CH4, and
This suggests a shift in the combustion process from the primary to the
their mixtures, Int. J. Hydrogen Energy 48 (2023) 24101–24112, https://doi.org/
secondary combustion zone as the inlet flow rate increases. The similar 10.1016/j.ijhydene.2023.03.189.
trend was also observed in NOx emissions, because NOx emissions are [6] A.E.E. Khalil, A.K. Gupta, Hydrogen addition effects on high intensity distributed
directly related to temperature. However, as the inlet flow rate is sub- combustion, Appl. Energy 104 (2013) 71–78, https://doi.org/10.1016/j.
apenergy.2012.11.004.
stantially increased, a diminishing rate of temperature growth is [7] Z. Li, X. Cheng, W. Wei, L. Qiu, H. Wu, Effects of hydrogen addition on laminar
observed, and NOx emissions reach a plateau. As the flow rate increases flame speeds of methane, ethane and propane: Experimental and numerical
beyond a certain threshold, the residence time of the fuel and oxidizer analysis, Int. J. Hydrogen Energy 42 (2017) 24055–24066, https://doi.org/
10.1016/j.ijhydene.2017.07.190.
mixture within the burner decreases, thereby restricting the extent to [8] S. de Persis, M. Idir, J. Molet, L. Pillier, Effect of hydrogen addition on NOx
which combustion can occur and limiting the achievable temperature formation in high-pressure counter-flow premixed CH4/air flames, Int. J.
rise. This behavior indicates that the burner is reaching its maximum Hydrogen Energy 44 (2019) 23484–23502, https://doi.org/10.1016/j.
ijhydene.2019.07.002.
capacity, with further increases in flow rate yielding minimal additional [9] A. Aniello, T. Poinsot, L. Selle, T. Schuller, Hydrogen substitution of natural-gas in
temperature rise and NOx emissions. An optimal inlet flow rate of 50.40 premixed burners and implications for blow-off and flashback limits, Int. J.
LPM for hydrogen in the fuel inlet and 200 LPM for air in the oxidizer Hydrogen Energy 47 (2022) 33067–33081, https://doi.org/10.1016/j.
ijhydene.2022.07.066.
inlet has been determined using the proposed weighted sum perfor- [10] A. Boretti, Hydrogen internal combustion engines to 2030, Int. J. Hydrogen Energy
mance indicator based on WSM. 45 (2020) 23692–23703, https://doi.org/10.1016/j.ijhydene.2020.06.022.
In this study, the fuel-to-air equivalence ratio was retained at 0.6, [11] P. Stuttaford, H. Rizkalla, Y. Chen, B. Copley, T. Faucett, Extended Turndown, Fuel
Flexible Gas Turbine Combustion System, in: Vol. 2 Combust. Fuels Emiss. Parts A
leaving the relationship between this ratio and temperature, and NOx
B, ASMEDC, 2010: pp. 483–492. doi: 10.1115/GT2010-22585.
emissions unexplored. As a recommendation for future research, an [12] A. Haj Ayed, K. Kusterer, H.-H.-W. Funke, J. Keinz, C. Striegan, D. Bohn,
investigation of the effects of varying fuel-to-air equivalence ratios on Improvement study for the dry-low-NOx hydrogen micromix combustion
both temperature distribution and NOx emissions is necessary. technology, Propuls. Power Res. 4 (2015) 132–140, https://doi.org/10.1016/j.
jppr.2015.07.003.
[13] A. Uemichi, K. Kouzaki, K. Warabi, K. Shimamura, M. Nishioka, Formation of ultra-
Declaration of competing interest lean comet-like flame in swirling hydrogen–air flow, Combust. Flame 196 (2018)
314–324, https://doi.org/10.1016/j.combustflame.2018.06.019.
[14] M. Dutka, M. Ditaranto, T. Løvås, NOx emissions and turbulent flow field in a
The authors declare that they have no known competing financial partially premixed bluff body burner with CH4 and H2 fuels, Int. J. Hydrogen
interests or personal relationships that could have appeared to influence Energy 41 (2016) 12397–12410, https://doi.org/10.1016/j.ijhydene.2016.05.154.
the work reported in this paper. [15] K. Souflas, P. Koutmos, On the non-reacting flow and mixing fields of an
axisymmetric disk stabilizer, under inlet mixture stratification and preheat, Exp.
Therm Fluid Sci. 99 (2018) 357–366, https://doi.org/10.1016/j.
Acknowledgements expthermflusci.2018.08.008.
[16] C.H. Cho, G.M. Baek, C.H. Sohn, J.H. Cho, H.S. Kim, A numerical approach to
reduction of NOx emission from swirl premix burner in a gas turbine combustor,
This study was supported by the Research Program funded by the Appl. Therm. Eng. 59 (2013) 454–463, https://doi.org/10.1016/j.
SeoulTech (Seoul National University of Science and Technology). applthermaleng.2013.06.004.
[17] C.E. Arrieta, A.M. García, A.A. Amell, Experimental study of the combustion of
natural gas and high-hydrogen content syngases in a radiant porous media burner,
Appendix A. Supplementary material
Int. J. Hydrogen Energy 42 (2017) 12669–12680, https://doi.org/10.1016/j.
ijhydene.2017.03.078.
Supplementary data to this article can be found online at https://doi. [18] C. Lu, L. Zhang, C. Cao, X. Chen, C. Xing, H. Shi, L. Liu, P. Qiu, The effects of N2
org/10.1016/j.applthermaleng.2024.124748. and steam dilution on NO emission for a H2/Air micromix flame, Int. J. Hydrogen
Energy 47 (2022) 27266–27278, https://doi.org/10.1016/j.ijhydene.2022.06.050.
[19] X. Liu, H. Aljabri, M. Silva, A.S. AlRamadan, M. Ben Houidi, E. Cenker, H.G. Im,
Hydrogen pre-chamber combustion at lean-burn conditions on a heavy-duty diesel

12
D.K. Le et al. Applied Thermal Engineering 259 (2025) 124748

engine: A computational study, Fuel 335 (2023) 127042, https://doi.org/10.1016/ [36] C. Meraner, T. Li, M. Ditaranto, T. Løvås, Effects of scaling laws on the combustion
j.fuel.2022.127042. and NO characteristics of hydrogen burners, Combust. Flame 214 (2020) 407–418,
[20] X. Liu, H. Aljabri, N. Panthi, A.S. AlRamadan, E. Cenker, A.T. Alshammari, https://doi.org/10.1016/j.combustflame.2020.01.010.
G. Magnotti, H.G. Im, Computational study of hydrogen engine combustion [37] X. Yang, W. Yang, S. Dong, H. Tan, Flame stability analysis of premixed hydrogen/
strategies: dual-fuel compression ignition with port- and direct-injection, pre- air mixtures in a swirl micro-combustor, Energy 209 (2020) 118495, https://doi.
chamber combustion, and spark-ignition, Fuel 350 (2023) 128801, https://doi. org/10.1016/j.energy.2020.118495.
org/10.1016/j.fuel.2023.128801. [38] G. Lopez-Ruiz, I. Alava, J.M. Blanco, Study on the feasibility of the micromix
[21] P. Barreiro, I. Alava, J.M. Blanco, G. Lopez-Ruiz, An assessment of the operating combustion principle in low NO H2 burners for domestic and industrial boilers: A
conditions of the micromix combustion principle for low NOx industrial hydrogen numerical approach, Energy 236 (2021) 121456, https://doi.org/10.1016/j.
burners: Numerical and experimental approach, Int. J. Hydrogen Energy 66 (2024) energy.2021.121456.
208–222, https://doi.org/10.1016/j.ijhydene.2024.04.052. [39] J. Li, Z. Zhao, A. Kazakov, F.L. Dryer, An updated comprehensive kinetic model of
[22] T. Poinsot, Denis Veynante, Theoretical and numerical combustion, AFNIL, 2022. hydrogen combustion, Int. J. Chem. Kinet. 36 (2004) 566–575, https://doi.org/
[23] P. Domingo, L. Vervisch, Recent developments in DNS of turbulent combustion, 10.1002/kin.20026.
Proc. Combust. Inst. 39 (2023) 2055–2076, https://doi.org/10.1016/j. [40] M.Ó. Conaire, H.J. Curran, J.M. Simmie, W.J. Pitz, C.K. Westbrook,
proci.2022.06.030. A comprehensive modeling study of hydrogen oxidation, Int. J. Chem. Kinet. 36
[24] T.-H. Shih, W.W. Liou, A. Shabbir, Z. Yang, J. Zhu, A new k-∊ eddy viscosity model (2004) 603–622, https://doi.org/10.1002/kin.20036.
for high reynolds number turbulent flows, Comput. Fluids 24 (1995) 227–238, [41] S.G. Davis, A.V. Joshi, H. Wang, F. Egolfopoulos, An optimized kinetic model of
https://doi.org/10.1016/0045-7930(94)00032-T. H2/CO combustion, Proc. Combust. Inst. 30 (2005) 1283–1292, https://doi.org/
[25] A.H. Ayed, K. Kusterer, H.-H.-W. Funke, J. Keinz, D. Bohn, CFD based exploration 10.1016/j.proci.2004.08.252.
of the dry-low-NO x hydrogen micromix combustion technology at increased [42] P. Saxena, F.A. Williams, Testing a small detailed chemical-kinetic mechanism for
energy densities, Propuls. Power Res. 6 (2017) 15–24, https://doi.org/10.1016/j. the combustion of hydrogen and carbon monoxide, Combust. Flame 145 (2006)
jppr.2017.01.005. 316–323, https://doi.org/10.1016/j.combustflame.2005.10.004.
[26] H.H.W. Funke, N. Beckmann, J. Keinz, S. Abanteriba, Comparison of numerical [43] H. Sun, S.I. Yang, G. Jomaas, C.K. Law, High-pressure laminar flame speeds and
combustion models for hydrogen and hydrogen-rich syngas applied for dry-low- kinetic modeling of carbon monoxide/hydrogen combustion, Proc. Combust. Inst.
nox-micromix-combustion, J. Eng. Gas Turbines Power 140 (2018), https://doi. 31 (2007) 439–446, https://doi.org/10.1016/j.proci.2006.07.193.
org/10.1115/1.4038882. [44] Gregory P. Smith, David M. Golden, Michael Frenklach, Nigel W. Moriarty, Boris
[27] M. Ferrarotti, W. De Paepe, A. Parente, Reactive structures and NOx emissions of Eiteneer, Mikhail Goldenberg, C. Thomas Bowman, Ronald K. Hanson, Soonho
methane/hydrogen mixtures in flameless combustion, Int. J. Hydrogen Energy 46 Song, J. William C. Gardiner, Vitali V. Lissianski, Zhiwei Qin, GRI-MECH 3.0, (n.
(2021) 34018–34045, https://doi.org/10.1016/j.ijhydene.2021.07.161. d.). http://www.me.berkeley.edu/gri_mech/ (accessed May 2, 2024).
[28] N. Schmidt, M. Müller, P. Preuster, L. Zigan, P. Wasserscheid, S. Will, Development [45] Z. Hong, D.F. Davidson, R.K. Hanson, An improved H2/O2 mechanism based on
and characterization of a low-NOx partially premixed hydrogen burner using recent shock tube/laser absorption measurements, Combust. Flame 158 (2011)
numerical simulation and flame diagnostics, Int. J. Hydrogen Energy 48 (2023) 633–644, https://doi.org/10.1016/j.combustflame.2010.10.002.
15709–15721, https://doi.org/10.1016/j.ijhydene.2023.01.012. [46] M.P. Burke, M. Chaos, Y. Ju, F.L. Dryer, S.J. Klippenstein, Comprehensive H 2 /O 2
[29] M. Kazemi, S. Brennan, V. Molkov, Numerical simulations of the critical diameter kinetic model for high-pressure combustion, Int. J. Chem. Kinet. 44 (2012)
and flame stability for hydrogen flames, Int. J. Hydrogen Energy 59 (2024) 444–474, https://doi.org/10.1002/kin.20603.
591–603, https://doi.org/10.1016/j.ijhydene.2024.02.039. [47] Y. Wang, S. Verhelst, Comparative analysis and optimisation of hydrogen
[30] ANSYS Fluent Theory Guide, (2023). https://www.ansys.com/ (accessed May 5, combustion mechanism for laminar burning velocity calculation in combustion
2024). engine modelling, Int. J.Hydrogen Energy 56 (2024) 880–893, https://doi.org/
[31] B. Magnussen, On the structure of turbulence and a generalized eddy dissipation 10.1016/j.ijhydene.2023.12.214.
concept for chemical reaction in turbulent flow, in: 19th Aerosp. Sci. Meet., [48] Z. Li, J. Wang, X. Li, Application of hydrogen mechanisms in combustion
American Institute of Aeronautics and Astronautics, Reston, Virigina, 1981. doi: simulation of DLR scramjet combustor and their effect on combustion performance,
10.2514/6.1981-42. Fuel 349 (2023) 128659, https://doi.org/10.1016/j.fuel.2023.128659.
[32] B.F. Magnussen, The eddy dissipation concept: A bridge between science and [49] M.A. Habib, G.A.Q. Abdulrahman, A.B.S. Alquaity, N.A.A. Qasem, Hydrogen
technology, in: ECCOMAS Themat. Conf. Comput. Combust., 2005: p. 24. combustion, production, and applications: A review, Alexandria Eng. J. 100 (2024)
[33] M. Skottene, K.E. Rian, A study of NOxNOx formation in hydrogen flames, Int. J. 182–207, https://doi.org/10.1016/j.aej.2024.05.030.
Hydrogen Energy 32 (2007) 3572–3585, https://doi.org/10.1016/j. [50] Y.A. Zeldovich, D. Frank-Kamenetskii, P. Sadovnikov, Oxidation of nitrogen in
ijhydene.2007.02.038. combustion, Publishing House of the Acad of Sciences of USSR, 1947.
[34] O. Cam, H. Yilmaz, S. Tangoz, I. Yilmaz, A numerical study on combustion and [51] A.L. Purohit, A. Nalbandyan, P.C. Malte, I.V. Novosselov, NNH mechanism in low-
emission characteristics of premixed hydrogen air flames, Int. J. Hydrogen Energy NOx hydrogen combustion: Experimental and numerical analysis of formation
42 (2017) 25801–25811, https://doi.org/10.1016/j.ijhydene.2017.07.017. pathways, Fuel 292 (2021) 120186, https://doi.org/10.1016/j.fuel.2021.120186.
[35] A. Cemal Benim, B. Pfeiffelmann, Validation of Combustion Models for Lifted [52] E. Triantaphyllou, Multi-criteria Decision Making Methods: A Comparative Study,
Hydrogen Flame, E3S Web Conf. 128 (2019) 01014, https://doi.org/10.1051/ Springer US, Boston, MA, 2000. doi: 10.1007/978-1-4757-3157-6.
e3sconf/201912801014.

13