See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245148561

Combustion characteristics of natural gas–hydrogen hybrid fuel turbulent

diffusion flame

Article in International Journal of Hydrogen Energy · March 2010

DOI: 10.1016/j.ijhydene.2009.12.049

CITATIONS READS

80 2,631

3 authors, including:

Sherihan El-Ghafour
Port Said University
9 PUBLICATIONS 158 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sherihan El-Ghafour on 10 October 2018.

The user has requested enhancement of the downloaded file.

 ARTICLE IN PRESS
international journal of hydrogen energy xxx (2010) 1–10

Available at www.sciencedirect.com

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

Combustion characteristics of natural gas–hydrogen hybrid

fuel turbulent diffusion flame

S.A.A. El-Ghafour*, A.H.E. El-dein, A.A.R. Aref

Mechanical Power Engineering Department, Faculty of Engineering, Suez Canal University, Port-Said, Egypt

article info abstract

Article history: Combustion characteristics of natural gas – hydrogen hybrid fuel were investigated
Received 27 June 2009 experimentally in a free jet turbulent diffusion flame flowing into a slow co-flowing air
Received in revised form stream. Experiments were carried out at a constant jet exit Reynolds number of 4000 and
9 December 2009 with a wide range of NG–H2 mixture concentrations, varied from 100%NG to 50%NG-50% H2
Accepted 9 December 2009 by volume. The effect of hydrogen addition on flame stability, flame length, flame struc-
Available online xxx ture, exhaust species concentration and pollutant emissions was conducted. Results
showed that, hydrogen addition sustains a progressive improvement in flame stability and
Keywords: reduction in flame length, especially for relatively high hydrogen concentrations.
Natural gas Hydrogen-enriched flames found to have a higher combustion temperatures and reactivity
Hydrogen than natural gas flame. Also, it was found that hydrogen addition to natural gas is an
Hybrid fuel ineffective strategy for NO and CO reduction in the studied range, while a significant
Free jet turbulent diffusion flame reduction in the %CO2 molar concentration by about 30% was achieved.
Hydrogen-enriched flame ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1. Introduction characteristics such as wide flammability range, high laminar

In today’s world, the problem of climatic change is one of the
serious drawbacks driving mainly from the emission of large
quantities of CO2, the inevitable product of fossil fuel
combustion [1]. This is not the only environmental hazard
from fossil fuels, but its emission of noxious gases such as
NOx, CO and SO2, results in critical environmental problems
throughout the world [2]. In addition, the declining of fossil
fuels supplies at an alarming rate exhibits the desire to
use it economically [3]. In view of these problems, it will
unquestionably require long-term changes in the fuels and
technologies that we use to meet our energy needs. Hydro-
carbon- hydrogen hybrid fuel has become as an attractive
option for reducing the dependency on the fossil fuels, and
provides a transition strategy to carbon free energy systems.
The choice of hydrogen as an essential participant in
hybrid fuels is due to: firstly, its superior combustion
flame speed, and low ignition energy, in addition to its high
molecular diffusivity [4,5]. Secondly, the environmental
benefits of its combustion, as there is no CO2, CO, SOx and UHC
emissions. Furthermore, hydrogen can be produced from
a variety of feedstocks; from fossil resources and renewable
resources [6]. In that manner, the utilization of hydrogen in
blended form would reduce the problems of storage and
flashback [4].
NG–H2 hybrid fuel, in particular, became the subject of
extensive research in recent years. A series of experimental
and numerical investigations has been done to present
a clearer understanding for its combustion characteristics.
Addition of hydrogen to natural gas combustion showed an
increase in flame adiabatic temperature, a reduction in flame
thickness and quenching distance and further enhances the
auto-ignition characteristics and the global rate of heat
release [7–10].

* Corresponding author. Tel.: þ20 194002162; fax: þ20 663400936.

E-mail address: sh.a.ghafour@gmail.com (S.A.A. El-Ghafour).
0360-3199/$ – see front matter ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2009.12.049

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
 ARTICLE IN PRESS
2 international journal of hydrogen energy xxx (2010) 1–10
Ilbas et al. [11] studied experimentally the influence of
hydrogen enrichment on the natural gas laminar flame
properties. They stated that increasing the hydrogen
percentage in mixture caused an increase in the resultant
burning velocity and also widening the flammability limits.
So, adding hydrogen to natural gas can extend the combustion
lean operation limit [12]. Such technique is currently recog-
nized as an effective approach to fuel economy and NOx
emission reduction. Improvement of the natural gas flame
stability behavior with hydrogen addition was a focal point of
study for pre-mixed [13] and diffusion [5,14,15] flames.
Kumar and Mishra [16] performed an experimental inves-
tigation on laminar LPG-H2 jet diffusion flame and observed
an insignificant variation in the visible flame length upto 20%
of hydrogen addition, but a subsequent reduction in the flame
length was observed for the higher hydrogen concentrations.
However, a continuous reduction in the NG-H2 turbulent
flame length with hydrogen addition was confirmed by
Choudhuri and Gollahalli [17].
It is clear that the flame structure characteristic were
deeply investigated. The height of the blue cone, a main
parameter depicts the flame structure, was studied in
different investigations [5,18]. It was observed that the height
of the blue cone decreased significantly when the amount of
hydrogen increased. Choudhuri and Gollahalli [4] explained in
detail the flame structure in terms of in-flame profiles of
temperature and composition at three axial locations of the
flame.
As reported by Choudhuri and Gollahalli [17] a significant
reduction in both CO emission and soot concentration with
hydrogen addition was noticed in the turbulent jet diffusion
flame. The tests of Burbano et al. [18] on two different
atmospheric burners also affirmed the same issue. On other
side, the results of Cozzi and Coghe [5] in a NG–H2 diffusion
swirled flame and Kumar and Mishra [16] in a LPG–H2 laminar
diffusion flame observed an increase in CO and soot
concentrations with the increase of hydrogen content in the
fuel stream.
Several attempts [4,5,19,20] have been made to predict the
change of NOx emission behavior in the NG-H2 flame. All of
these researches agreed upon that NOx emission increases
with hydrogen addition. However, they disputed about the
formation mechanism responsible for the increase of NOx
emission. Rortveit and Hustad [20] conducted a numerical
and experimental study of counterflow NG/H2 – air partially
premixed flame. They evaluated the contributions of the
thermal and the prompt NOx mechanisms in the total NOx
formation and found that the prompt NOx route was the
main contributor. In contrast, the results of Naha and
Aggarwal [19] confirmed that the thermal NOx was more
dominant than the prompt NOx for the same flame
configuration.
Although the above survey manifests many substantial
efforts that have been devoted to investigate different char-
acteristics of NG–H2 hybrid fuel, a comprehensive under-
standing for its behaviors under the different operating
conditions has not been yet consummated. Therefore, the
author present a closer study to investigate experimentally
fundamental combustion characteristics of NG–H2 hybrid fuel
at a range of fuel jet Reynolds number varies from 3000 to 6500
Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
is carried out, available in the thesis by El-Ghafour [21].
However, results in this paper selected only at a jet Reynolds
number of about 4000. The experimental procedures were
conducted in a simple free jet turbulent diffusion flame
flowing into a slow co-flowing air stream.
2. Experimental setup and procedures
2.1. Experimental setup
Fig. 1(a) summarizes the experimental facility used in this
study. The essential features of the setup are a combustion
chamber, fuel and co-flow air supply systems, transverse
mechanism, and measuring instruments.
The combustion chamber is similar to that described by
Gollahalli and Wright [22], but with some modifications to
match the objectives of the current experiments. The
chamber is constructed of a 3.0 mm steel sheet and fabricated
as a cuboid of a square cross section, 0.6 m  0.6 m, and 1.24 m
height. This chamber is erected vertically, to sustain an axial
symmetry as far as possible, on a supporting frame and
equipped with rectangular windows of dimensions
0.2 m  1 m on all of its four side-walls. Three of these
windows are fitted with high temperature air-cooled Pyrex
glass plates of 3 mm thickness to allow visual observation of
the test events and for direct photography. A steel sheet,
provided with a longitudinal aperture, is fixed through the
fourth window for introducing the intrusive instruments. At
the top of the chamber, an exhaust duct of 0.85 m height is
connected. The fuel jet is discharged vertically upward along
the centerline of the combustor from a circular stainless steel
burner of 3.5 mm inner diameter and 0.75 m long, which is
long enough to achieve fully developed flow [4]. The burner is
projected 100 mm above the combustor floor. At the
combustor base, a vertical steel flow straightener tube, of
0.2 m inner diameter and 0.5 m long, is installed to ensure
a uniform velocity of the co-flowing air stream.
The co-flowing air is supplied by a centrifugal blower of
0.37 kW. An arrangement of a 75 mm thick layer of small glass
balls (9 mm diameter), a coarse wire mesh, and a fine wire
mesh is placed in the straightener tube in series. The details of
the co-flowing air section are shown in Fig. 1(c). The unifor-
mity of the air flow, at the chamber circular opening exit, is
verified by using a hot wire anemometer.
The fuel supply system is composed of separated natural
gas and hydrogen feeding lines, mixing section, and manifold
to the chamber burner. Commercial natural gas (mainly
composed of 91.43% CH4, 4.1% C2H6, 0.99% C3H8, 0.33%C4H10,
0.1%C5H12, 0.52% CO2, 2.53%N2), and hydrogen (99.94% purity)
are used. A hydrogen flashback arrestor is positioned inline
for safety reasons. The fuels flow rates are measured by two
corrected rotameters. Hydrogen and natural gas are led to the
mixing section where proper mixing is achieved, as shown
schematically in Fig. 1(b). Inside this section, hydrogen is
injected axially through a single hole nozzle, while natural gas
is injected radially into the hydrogen stream, immediately
downstream the hydrogen nozzle, through four holes equally
distributed around a circumference perpendicular to the
nozzle axis. Finally, a throttle part was employed to ensure
 ARTICLE IN PRESS
international journal of hydrogen energy xxx (2010) 1–10 3

Fig. 1 – Schematic of (a) Experimental setup, (b) Mixing section, dimensions in (mm) and (c) Co-flowing air section.

more effective mixing in the mixing section. This arrange-
ment ensures a good condition for swift mixing process
between two different gases [23]. The mixture is then intro-
duced into a long hose, with length of more than 150 times its
hydraulic diameter, to ensure homogenous mixture before
entering the fuel burner.
A fine wire Platinum–Platinum/13% Rhodium, of 0.24 mm
wire diameter and 0.45 mm bead diameter, exposed junction
thermocouple is used for measuring local temperatures. The
output temperature is sampled at a rate of 500 Hz and time-
averaged for 7.5 s, after reaching the stabilized reading, using
a PC-based high speed data acquisition system (NI model USB-
1961) with a windows-based DAQ software. A temperature
correction is made for radiation and conduction heat losses by
the use of Bell technique [24].
The species concentrations of O2, CO and NO are
measured by a multi-gas analyzer (Testo model 350), which
consists of a bundle of three measuring cells and equipped
with a CO2 calculation program. Gas samples from inside
the flame are withdrawn through an un-cooled narrow tip-
ped quartz probe of 0.8 mm orifice diameter. For exhaust
products measurement, a Pyrex glass cone is mounted
above the visible flame, and aligned with the burner axis.
The chosen collector location, depending on the visible
flame length, is sufficiently far downstream of the reaction
zone to ensure that all the relevant reactions have achieved
a steady state.
The minimum sampling time required to obtain nearly
steady gas concentrations for each reading, within 4%
deviation from its mean value, is 150 s.
2.2. Test conditions and procedure
Test conditions offer a specific range of hybrid fuel compo-
sitions, starts with pure NG (0% H2); as the baseline test fuel,
and by a hydrogen increment of 10% (by volume), the
mixture composition gradually changes to 50%NG–50% H2.

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
 ARTICLE IN PRESS
4 international journal of hydrogen energy xxx (2010) 1–10

Table 1 – Investigated test conditions (Rej [ 4000, dj [ 3.5 mm, ua [ 0.4 m/s).
Test No. Fuel hydrogen rf(kg/m3)a mf(Pa.s)a  106 Qf(m3/s)  103 mj(m/s)b P(kW)c
content, vol%

1 0 0.7037 11.678 0.181 18.802 5.989

2 10 0.6414 11.359 0.193 20.070 5.926
3 20 0.5791 11.044 0.208 21.619 5.90
4 30 0.5167 10.735 0.227 23.542 5.917
5 40 0.4544 10.432 0.250 26.016 6.147
6 50 0.3921 10.137 0.282 29.30 6.332

a Based on cold fuel conditions of 304̊K, 1 atm.

b Jet (burner) exit velocity.
c Fuel thermal load.

Throughout the tests, the jet exit Reynolds number, Rej, is

held constant at about 4000.
The fuel volumetric flow rate,Qf, of NG–H2 hybrid at the
selected jet Reynolds number can be calculated as follow [25]:
" #
P Rej  mf  dj
Qf ¼ (1)
4 rf
where rf and mf are the mixture density and dynamic viscosity,
respectively. mf can be computed by the correlation suggested
by Nabizadeh and Mayinger [26] for NG–H2 mixtures. Table 1
summarizes the laboratory test conditions, along with its
physical properties.
Measurements are reported for flame stability (liftoff and
blowout behaviors), visible flame appearance, flame structure,
exhaust CO2 and O2 concentrations as well as NO and CO
pollutant emission indices.
In the present study the co-flowing air velocity, ua, at the
circular opening exit, is held constant at 0.4 m/s (Rea ¼ 5300);
except for the stability tests. This velocity was sufficient for
stable combustion under all current tested conditions. Stability
tests are studied by incrementally increasing the co-flowing
stream velocity, while maintaining the fuel flow rate constant,
until the flame is visually observed to liftoff or blowout.
Direct flame photography, with exposure time of 0.5 s, was
used to document the changes in appearance and flame
length with changes in the fuel composition. For each case,
twenty exposures are taken. Subsequently, average flame
length from these photographs is estimated. The flame length
is defined, experimentally, as the average distance between
the jet exit plane of the fuel burner and the flame tip where the
flame luminosity is no longer visible to eye.
For the thermal flame structure tests, thermocouple is
mounted on a two-dimensional traversing mechanism
coupled with a steeper motor that has an accuracy of 0.05 mm
radially and 0.5 mm axially. The mechanism allows smooth
radial movement of the thermocouple through the flame, with
an interval of 1 mm, from r ¼ 0–40 mm; and axial movements
with an interval of 10 mm, from X ¼ 0–200 mm, and with an
interval of 15 mm, from X ¼ 200–600 mm. An exact 1880 data
point are obtained and plotted in color scale contour plots.
2.3. Uncertainty estimates
To ensure experimental repeatability for the stability and
flame length measurements, each test is carried out three
Table 2 – Estimated uncertainties, %, for different
measuring parameters.
Parameter Uncertainty, %
Temperature 1.2
Fuel Flow rate 2.0
Air Velocity 0.5
CO Concentration 3.1
NO Concentration 3.9
O2 Concentration 3.0
times and the averaged data were reported and employed to
conduct the uncertainty analysis. Uncertainty of the liftoff
and blowout limits measurements are estimated to be about
1.4% and 2.2%, respectively. Flame length and fluctuation
measurements uncertainty are within 5.1% and 4.2%,
respectively. The estimated uncertainties for the other
measuring components are reported in Table 2. Uncertainty of
these components is estimated by combining the systematic
and random errors. Further details of the uncertainty analysis
can be found in El-Ghafour [21].
3. Results and discussion
3.1. Liftoff and blowout behaviors
The phenomena of liftoff and blowout were investigated by
many researchers. Many theories were developed to give an
explanation for these phenomena [27]. The present study
relies on one of these theories as the main reference for
results analysis. This theory presumes that the imbalance of
the chemical reaction rate S2l =a, and the strain rate,U/d, is the
primary cause of the flame liftoff and blowout, where Sl is the
maximum laminar burning velocity,a is the thermal dif-
fusivity,U is the local centerline velocity, and d is the local jet
width.
The high laminar burning velocity of hydrogen explains
the greater stability of this fuel, its speed is more than five
times that of CH4 [28]. Thus, increasing the %H2 content in the
hybrid fuel increases the flame reaction rate and enhances the
stability behaviors.
Effects of hydrogen content on the value of co-flowing air
velocity at the liftoff and blowout conditions are displayed in

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
 ARTICLE IN PRESS
international journal of hydrogen energy xxx (2010) 1–10
Fig. 2 – Measured flames liftoff and blowout limits of NG-H2
hybrid fuel as a function of co-flowing air stream velocity.
The dashed lines represent a polynomial fit to all of the
data.
Fig. 2. It is found that the co-flowing air velocity at the liftoff and
blowout conditions increases gradually as the hydrogen
content is increased till 45% H2. Beyond this value, the co-
flowing air velocity increases sharply with further increase of
the hydrogen content. It can be noticed from Fig. 2 that the
change of hydrogen addition from 0% to 50% raises the blowout
co-flowing velocity by approximately 2.8 times. Although,
increasing the co-flowing velocity directly increases the strain
rate,U/d, the superior combustion characteristics of hydrogen
become more dominant on the stability behavior [15].
Moreover, the stability diagram can be divided into two
distinct regimes: (i) first regime, till nearly 45% H2 content, the
attached flames become lifted before the blowout, and (ii)
second regime, beyond 45% H2, the attached flames blowout
suddenly without converting to the lifted configuration
(blowout as attached flames).
3.2. Visible flame appearance
In this part; the flames formal variations as a result of
increasing hydrogen content in the hybrid fuel stream were
investigated. Color photographs of the tested flames are
available in El-Ghafour [21].
Fig. 3 – Variation of the normalized flame length, Lf/dj, as
a function of %H2 concentration in the hybrid fuel. Error
bars represent the visual fluctuations in the normalized
flame lengths.
Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
5
Initially, the pure natural gas flame appears as a lifted
flame. Based on the above discussion, an expected enhance-
ment in the flame stability with hydrogen addition is noticed
and confirmed in Fig. 4 as a conversion from a fully lifted to
a burner attached configuration. Furthermore, the presence of
"a partially premixing level" found to be an effective param-
eter in the tested flames comparison. The effect of this level
extends over the entire combustion characteristics and it will
be exposed to its effects in turn below. The second observation
concerns with the flames colors, which is essentially attrib-
uted to the effect of partial premixing degree up to 40% of
hydrogen addition. The natural gas flame firstly is a blue in
color, with a small luminous yellow region at the flame tip. A
gradual spreading for the yellow color flame, triumphed over
the blue color, is observed with the increase of % H2 fraction.
Until the hydrogen fraction reaches 40%, the entire flame
becomes yellow except the near burner region (free soot
region) still blue in color. The entity of the yellow colored
regions in the flames is attributed to the soot formation,
growth and oxidation processes [4]. Whereas, a faintly emer-
gence of a pale red color, for the 50% H2 content case, is
occurred due to both H2O emission and the small impurities
contained in the H2 fuel [5].
Furthermore, the effect of higher molecular diffusivity of
hydrogen appears clearly as an increase in the visible blue
zone width for the near burner region, when hydrogen addi-
tion is varied from 40% to 50%.
3.2.1. Visible flame length
Fig. 3 illustrates the variation of the normalized flame length,
Lf/dj, with the % H2 content. From this figure, it can be seen
that the flame length slightly increases up to 20% of H2 content
then decreases markedly at higher hydrogen concentrations.
Generally, the flame length decreases as the hydrogen
concentration increases in the NG-H2 hybrid fuel [4]. This can
be attributed to the increase in the radical pool that accom-
panies the H2 addition, such as H and OH radicals, which
consequently enhances the combustion rate and thereby
reduces the overall flame length [17]. In addition, the
hydrogen low molecular weight and rapid radial molecular
diffusion causes an increase in the flame width, which is
conversely proportional to the flame length. However, in the
present work, at Re ¼ 4000, the partially premixing level also
found to highly affect the overall flame length.
The obtained results may be explained in the following
way; the initial increase in the flame length is imputed to the
decreasing of the liftoff height, accompanied by decreasing
in the vertical flame fluctuations (see Fig. 3), which causes an
elongation in the flame length with about 2.5% for hydrogen
addition ranging from 0 to 20%. Also, it is well known that the
flow in the high temperature regions is accelerated upwardly
due to the buoyancy so that the flame is somewhat elon-
gated. While, the dominance of H2 combustion characteris-
tics may become the main responsible for the subsequent
reduction in the flame length. It can be noticed from Fig. 3
that the percentage reduction in the normalized flame length
is about 12% as the H2 concentration changes from 0 to 50%.
This result depicts a good agreement with that available in
the existing literature (Kumar and Mishra [16]) for LPG-H2
hybrid fuel.
 ARTICLE IN PRESS
6 international journal of hydrogen energy xxx (2010) 1–10

Fig. 4 – Comparison of measured flames temperature distributions ˚[K] (left) with radial temperature profiles at three axial
locations (right): (1) near-burner, (2) mid-flame, and (3) far-burner.

3.3. Flame structure
3.3.1. Temperature distribution
Fig. 4 presents the contours of measured flames temperature
distribution in the left half with radial profiles of the measured
temperature at three axial locations; near-burner (0.25 Lf),
mid-flame (0.5 Lf) and far-burner (0.75 Lf), in the right half.
It can be seen from Fig. 4 the changes occur in some
features of the flame appearance, mentioned above, such as
the transition in the flame configuration and the change in the
flame length. Moreover, the maximum temperature zone in
the baseline flame (0%H2) is clearly disappeared. This can be
explicated by the higher partial premixing level, which results
in a larger air entrainment and, consequently, a more rela-
tively cooler flame structure.
It can be demonstrated also that there are gradual broad-
ened in the high temperature zone as well as a continuous
movement of thermal flame envelope in the upstream direc-
tion; closer to the burner exit, with H2 addition in the fuel
stream. This is due to the hydrogen combustion characteris-
tics, high adiabatic temperature and reactivity [25].
A single hump structure with off-axis peak is noticed in the
‘‘near-burner region’’. The single hump structure occurs
because of the gas phase reaction is dominant in the near-
burner region, although the pyrolysis reaction leading to soot
formation has already begun [29]. At the stoichmetric contour,

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
 ARTICLE IN PRESS
international journal of hydrogen energy xxx (2010) 1–10 7

the 0%H2 flame has a peak temperature of 1483 K, but with the
increase of H2 fraction in the fuel stream, the peak tempera-
ture starts to increase due to the increase in the adiabatic
flame temperature. In addition, the stoichmetric contours are
shifted towards the rich side, indicating the increase of the
overall reaction rate with hydrogen addition.
As the jet spreads, the radial temperature profiles become
flatter at the downstream locations. However the off-axis
peaks are still recognizable. This can be clarified from the
comparison of the near-burner profiles with mid-flame and
far-burner profiles, as seen in the figure.
Generally, for the ‘‘mid-flame region’’ the flame stoichmetric
contour moves inward to the burner axis, as compared to that
for the near-burner region [30]. For the relatively low hydrogen
concentration flames, near-burner and mid-flame temperature
profiles have flatter peak temperature regions. It is a strong
indication of better mixing due to the presence of partially
premixing zone. As pointed out earlier, hydrogen addition to the
fuel stream up to 40% reduces the degree of partial premixing
and thereby increases the flame tendency to soot formation.
This is quite understandable that the mid-flame zone has the
highest soot concentrations [4], since this is the area for soot
particle inception and growth. So a relative decrease in the
measured flame temperatures is noticed due to larger radiative
heat loss for the high hydrogen concentration cases (more than
30%) when compared to the low concentration cases. Fig. 5 – Measured radial NO concentrations profiles for
Temperature profile of the ‘‘Far-burner region’’ is always different %H2 volume fractions in the hybrid fuel at three
characterized by a single peak close to the burner axis [4]. axial locations: (i) near-burner, (ii) mid-flame, and (iii) far-
Since the heat release in the far-burner region is controlled by burner.
heterogeneous combustion of soot [29], an increase in the
peak flame temperatures with hydrogen addition is recog-
nized. This seems to be due to the expected increase in the
soot oxidation rate with hydrogen addition due to the increase
of OH radical concentration (the soot dominant oxidizer) [30]. on the hydrogen rich radicals, H and OH. Although, the
increase of % H2 addition may cause an expected reduction in
3.3.2. In-flame NO concentrations C/H ratio and therefore flame CH radicals, some recent studies
In natural gas and natural gas-hydrogen hybrid fuel had another opinion [20,32]. Rortveit and Hustad [20]
combustion, the formation of NO can be attributed to two presumed experimentally the enhancement of the prompt
main chemical routes, which are the Zel’dovich thermal and rate by H2 addition to NG up to 50% H2 fraction through steps
Fenimore prompt NO mechanisms [31]. (A) and (B).
As it is already seen in Fig. 4, the flames temperatures are Also, it should be mentioned that the increase of hydrogen
considerably lower than 1800 K, so the prompt Fenimore addition in the hybrid fuel stream might increase the thermal
mechanism becomes significant in comparison to the thermal NO contribution in the total NO production [33], especially, for
mechanism [28]. the relatively high hydrogen concentrations cases (e.g., 40%
It is well known that CH radical, the main precursor for and 50% H2). This can be relied on the increase in: (i) the peak
Fenimore mechanism, is produced in the reaction zone and flame temperature, which rises by about 9% with hydrogen
immediately leads to the formation of HCN through the variation from 0 to 50%, (ii) the residence time of combustion
reaction CH þ N24 HCN þ N and the subsequent conversion gases in the high temperature zone, this may be clarified from
of HCN to NO occurs in the availability of OH and O radicals. the growth of the high temperature region with %H2 addition
Moreover, the C2- species is considered the major source of CH (see Fig. 4) and (iii) the availability of OH radicals. According to
radicals through steps [32]: that the increase of the in-flame NO concentrations with
hydrogen addition is justifiable. This can be clearly confirmed
from Fig. 5, which exhibits the radial concentration profiles of
CH2þH 4 CH þ H2 (A) NO at the near-burner, mid-flame, and far-burner locations,
respectively, for selected tested flames of: 0%, 20%, 40% and
50% H2 fraction in the fuel. A single hump structure with off-
CH2þOH 4 CH þ H2O (B) axis peak is noticed in the near-burner region. However, for
downstream locations, the peak regions become flatter,
The above reactions, implies that the execution of NO compared to near-burner location, while its off-axis peak is
formation by Fenimore mechanism is significantly dependent still recognizable.

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
 ARTICLE IN PRESS
8 international journal of hydrogen energy xxx (2010) 1–10

The peak NO concentrations in the three studied positions

are located inward of the corresponding peak temperature;
i.e., in the fuel rich side of the flame sheet. Since the NO
formation can be correlated to the CH radical concentration
profiles which have peaks on the rich side of the fuel–oxidizer
interface [34].

3.4. Pollutant emission indices

The amount of NO and CO emissions is conventionally

quantified by the emission index ‘EIi’, which is generally
defined as the mass of pollutant emitted per unit mass of fuel
consumed. Species ‘EIi’ can be obtained from the measured
species molar concentration using the following equation [28]:
  !
xi n$Mwi
EIi ¼
xCO2 þ xCO Mwf
where Mw and x denote the molecular weight and mole frac-
tion, respectively, n is the number of carbon moles in a mole of
fuel, and the subscripts i and f represent species ‘i’ and fuel,
respectively.
3.4.1. NO emission index (EINO)
Based on the above discussion, the amount of NO emission
and, consequently, EINO level is assuredly increased with %H2
addition, as can be observed in Fig. 6.
All references studied the different configurations of NG-H2
diffusion flames agree unanimously on that; any hydrogen
addition to natural gas is an ineffective strategy for NO
reduction, although, the disparity in the reasons for which the
researchers had ascribed these increase.
3.4.2. CO emission index (EICO)
In the present research, the change in the EICO level with H2
addition in the hybrid fuel stream is reigned by two main
parameters: (i) flame residence time, tr ¼ L3f =uj $d2j , [35] and (ii)
rate of carbon input. According to the competition between
these parameters, the trend of EICO can be specified.
As illustrated in Fig. 6, the EICO level increases rapidly with
% H2 fraction, reaches the maximum value at 30% H2, and
then decreases faintly with the further increase of % H2.
Although, there is an increase in the flame length as H2
content varies from 0 to 20%, a relatively higher increase in uj
values with hydrogen is also noticed, as listed in Table 1.
Fig. 6 – Measured NO and CO emission indices (EINO, EICO)
as a function of %H2 concentration in the hybrid fuel.
Fig. 7 – Measured CO2 and O2 molar concentrations as
a function of %H2 concentration in the hybrid fuel.
(2)
Therefore, an expected decrease in the tr with %H2 fraction
takes place, especially for the short flame length cases. It is
well known that the reduction in tr do not allow for the
completion of CO oxidation process [28]. This explains the
dramatic increase of EICO level in the first segment of the
trend. In this context, it is worth mentioning that the change
in the partially premixing level with H2 addition from 0 to 30%
also plays a significant role in lowering the degree of CO
oxidation [19]. Whereas, in the second segment of trend;
beyond 30% H2, the change in EICO level submits to the effect
of carbon input rate reduction.
These results are consistent with those reported by Kumar
and Mishra [16]. Also, there is a good qualitative agreement
between the present results and those denoted by Cozzi and
Coghe [5]. Whereas, Choudhuri and Gollahalli [4] confirmed
that a continuous reduction in EICO level with H2 addition was
achieved.
3.5. Exhaust CO2 and O2 concentrations
Fig. 7 exhibits the variations of CO2 and O2 molar concentra-
tions in the exhaust product as a function of %H2 content in
the fuel stream.
The monotonic reduction of %CO2 with the fuel hydrogen
content is compatible with the dramatic increase of the CO
concentration, even when the H2 fraction is on the order of
30%, While, the continuation for this reduction is believed to be
due to the effect of carbon input diminution, as stated above.
From the figure, it is also clear that the 50% H2 flame can
generate approximately 30% less CO2 as compared to the pure
natural gas flame. This decrease in the % CO2 concentration is
the most significant gain for the NG-H2 hybrid fuel application.
An expected increase in the %O2 concentration is noticed
for H2 fraction from 0 to 40%. This behavior is mostly related to
the degree of partially premixing level, which acts upon the
flame air inflow and the mixing rate. Having reached to the
attached configuration, the effect of partial premixing
vanishes and another parameter dominates. This parameter
is the hydrogen higher diffusion coefficient, which is the main
responsible for % O2 concentration reduction in the 50%H2
flame as compared to 40%. An attracting remark should be
mentioned here; the higher values of % O2 concentration are
ascribed to the non-usage of any system for air entrainment
and mixing improvement.

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
 ARTICLE IN PRESS
international journal of hydrogen energy xxx (2010) 1–10
4. Conclusion
The combustion characteristics of NG-H2 hybrid fuel with
different compositions in a free jet turbulent diffusion flame
(Re ¼ 4000) with a slow co-flowing air stream were investi-
gated. By the addition of hydrogen to the fuel stream, in the
range of 0–50% by volume, the following can be concluded:
 An increase in the flame stability appeared as a significant
enhancement in the flame liftoff and blowout limits, which
can be attributed to the high burning velocity of hydrogen fuel.
 An obvious transition from a fully lifted to a burner attached
configuration was occurred, accompanied by an evident
change in the flame shape. This change was essentially
imputed to the effect of the partial premixing level upto 40%
of hydrogen addition, followed by a clear emergence for the
hydrogen combustion characteristics.
 A slight increase in the visible flame length was occurred
initially, with hydrogen addition upto 20%, then, an expec-
ted decrease was observed. The decrease in the liftoff height
with hydrogen addition explained the first increase,
whereas the higher hydrogen reactivity explained the
subsequent decrease.
 A continuous increase in flame temperatures, a broadened
in the high temperature zones as well as a gradual shifting
of the reaction zone in the upstream direction was observed,
which was related to the hydrogen higher reactivity and
adiabatic temperature
 The increase in the NO formation by the prompt mecha-
nism, as the main contributor, demonstrated the observed
increase in the in-flame NO and the value of EINO.
 Initially, EICO level increased with the hydrogen fraction
and reached a maximum value at 30% of H2 fraction, then,
a continuous reduction was observed. The decrease in the
flame residence time and the degree of partial premixing
explained the initial increase, while the reduction in carbon
input rate explained the subsequent decrease.
 A significant reduction in the %CO2 molar concentration by
about 30% can be achieved by adding hydrogen in the order
of 50%.
5. Recommendation for future work
Because of the diffusion flames are highly susceptible to scale
effects, future investigation of the burner diameter, and co-
flowing air velocity effects should be done. Also, more detailed
measurements on the in-flame NO and the exhaust NO
emission will be considered; to accurately decide the contri-
bution of each NO formation mechanisms. In addition, the
measurements of the soot concentrations and UHC emissions
will be a focal point of study in the future work.
references
[1] Genovese A and Ragona R. Public transport and CO2
emission: a comparative assessment between conventional
and innovative vehicles, In: proceeding of the 6th
Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
9
international conference of technologies and combustion for
a clean environment. Porto, Portugal, 1; 2003. p. 475–80.
[2] Momirlan M, Veziroglu TN. The properties of hydrogen as
fuel tomorrow in sustainable energy system for a cleaner
planet. Int J Hydrogen Energy 2005;30:795–802.
[3] Das LM, Gulati R, Gupta PK. A Comparative evaluation of the
performance characteristics of a spark ignition engine using
hydrogen and compressed natural gas as alternative fuels.
Int J Hydrogen Energy 2000;25:783–93.
[4] Choudhuri AR, Gollahalli SR. Combustion characteristics of
hydrogen-hydrocarbon hybrid fuels. Int J Hydrogen Energy
2000;25:451–62.
[5] Cozzi F, Coghe A. Behavior of hydrogen-enriched non-
premixed swirled natural gas flames. Int J Hydrogen Energy
2006;31:669–77.
[6] Akande A.J. Production of hydrogen by reforming of crude
ethanol. ph.D Dissertation, University of Saskatchewan
Saskatoon, Saskatchewan; 2005.
[7] Lafay Y, Renou B, Cabot G, Boukhalfa M. Experimental and
numerical investigation of the effect of H2 enrichment on
laminar methane–air flame thickness. Combust Flame 2008;
153:540–61.
[8] Fukuda M, Korematsu K, Sakamoto M. On quenching
distance of mixtures of methane and hydrogen with air.
JSME 1981;24(193):1192–7.
[9] Dülger Z. Adiabatic flame temperature and product
composition for lean combustion of hydrogen-methane
combination. ASME J.E Power 1994;2:707–12.
[10] Safta C, Madnia C. Autoignition and structure of
nonpremixed CH4/H2 flames: detailed and reduced kinetic
models. Combust Flame 2006;144:64–73.
[11] Ilbas M, Crayford A, Yılmaz I, Bowen P, Syred N. Laminar-
burning velocities of hydrogen–air and hydrogen–methane–
air mixtures: an experimental study. Int J Hydrogen Energy
2006;31:1768–79.
[12] Tseng C.J. Premixed combustion of hydrogen/methane
mixtures in a porous medium burner. In: proceeding of the
36th intersociety energy conversion engineering conference.
Savannah, GA, United States, 1; 2001. p. 589–94.
[13] Shy S, Chen Y, Yang C, Liu C, Huang C. Effects of H2 or CO2
addition, Equivalence ratio, and turbulent straining on
turbulent burning Velocities for lean premixed Methane
combustion. Combust Flame 2008;153:510–24.
[14] Choudhuri AR and Gollahalli SR. Effects of isothermal mixing
process on the stability of hydrogen-hydrocarbon blend
turbulent jet diffusion flame. In: proceeding of 2000 international
joint power generation conference, Florida. p. 1–11.
[15] Karbasi M, Wierzba I. The effects of hydrogen addition on the
stability limits of Methane jet diffusion flame. Int J Hydrogen
Energy 1998;23(No. 2):123–9.
[16] Kumar P, Mishra DP. Experimental investigation of laminar LPG-
H2 jet diffusion flame. Int J Hydrogen Energy 2008;33:225–31.
[17] Choudhuri AR, Gollahalli SR. Characteristics of hydrogen-
hydrocarbon composite fuel turbulent jet flames.
Int J Hydrogen Energy 2003;28:445–54.
[18] Burbano HJ, Amell AA, Garcia JM. Effects of hydrogen
addition to Methane on the flame structure and CO
emissions in atmospheric burners. Int J Hydrogen Energy
2008;33:3410–5.
[19] Naha S, Aggarwal SK. Fuel effects on NOx emissions in
partially premixed flames. Combust Flame 2004;139:90–105.
[20] Rortveit G and Hustad J. NOx formation in diluted CH4/H2
counterflow diffusion flames. In: proceeding of the 6th
international conference of technologies and combustion for
a clean environment. Porto, Portugal, 1; 2003.p. 339–44.
[21] El-Ghafour SA. Pollutant emissions control utilizing hybrid
fuel and air nozzle-cascading techniques, M.Sc. Thesis, Suez
Canal Univ.; 2009.
 ARTICLE IN PRESS
10 international journal of hydrogen energy xxx (2010) 1–10

[22] Gollhalli SR, Wright EA. Emission control from diffusion
flames. Houston, Texas, USA: International Joint Power
Generation Conference and Exposition; 1996.
[23] Davies AC. Welding. 10th ed. Cambridge University Press;
1996.
[24] Bradley D, Mathews KJ. Measurements of high gas
temperature with fine wire thermocouple. J Mech Eng 1968;
10(No 4):299–305.
[25] Choudhuri A.R. Experimental and numerical investigation
on hydrogen-hydrocarbon hybrid fuel combustion, ph.
d dissertation, School of aerospace and mechanical
engineering, The University of Oklahoma, Norman,
Oklahoma; 2000.
[26] Nabizadeh H. Mayinger F. Viscosity of binary mixtures of
hydrogen and natural gas (hythane) in the gaseous phase. In:
15th european conference on thermophysical properties,
Germany: Würzburg; 1999.p. 19–30.
[27] Feikema D, Chen RH, Driscoll JF. Blowout of nonpremixed
flames:maximum coaxial air velocities achievable, with and
without swirl. Combust Flame 1991;86:347–58.
[28] Turns SR. An introduction to combustion: concepts and
applications. New York, NY: McGraw-Hill, Inc.; 1996.
[29] Kamal A, Gollahalli S. Effects of jet reynolds number on the
performance of axisymmetric and nonaxisymmetric gas
burner flames. J Energy Resour Technol 2001;123:167–72.
[30] Qubbaj AR. and Gollhalli SR. Effects of induced air-flow on in-
flame concentrations of radicals. AIAA-0782; 2001. p. 1–12.
[31] Ilbas M, Yilmaz I, Kaplan Y. Investigation of hydrogen and
hydrogen-hydrocarbon composite fuel combustion and NOx
emission characteristics in a model combustor.
Int J Hydrogen Energy 2005;30(10):1139–47.
[32] Li SC, Williams FA. NOx formation in two-STage methane air
flames. Combust Flame 1999;118:399–414.
[33] Stroemman AH, Fossum M and Hustad J.E. Nitric oxide
formation in jet diffusion flames of CH4/H2/CO Mixtures. In:
proceeding of the 6th international conference of
technologies and combustion for a clean environment, Porto,
Portugal, 1; 2001. p. 325–30.
[34] Choudhuri AR, Gollahalli SR. Intermediate radical
concentrations in hydrogen-natural gas blended fuel jet
flames. Int J Hydrogen Energy 2004;30. 1303–1302.
[35] Sanders JP, Chen JY, Gokalp I. Flamelet-based modeling of
NO formation in turbulent hydrogen jet diffusion flames.
Combust Flame 1997;111:1–15.

Please cite this article in press as: El-Ghafour SAA, et al., Combustion characteristics of natural gas–hydrogen hybrid fuel
turbulent diffusion flame, International Journal of Hydrogen Energy (2010), doi:10.1016/j.ijhydene.2009.12.049
View publication stats