processes

Article
Fundamental Study on Hydrogen Low-NOx Combustion Using
Exhaust Gas Self-Recirculation
Kenta Kikuchi * , Tsukasa Hori and Fumiteru Akamatsu

Department of Mechanical Engineering, Osaka University, Osaka 5650871, Japan;

[email protected] (T.H.); [email protected] (F.A.)
* Correspondence: [email protected]; Tel.: +81-3-3803-9804

Abstract: Hydrogen is expected to be a next-generation energy source that does not emit carbon
dioxide, but when used as a fuel, the issue is the increase in the amount of NOx that is caused by
the increase in flame temperature. In this study, we experimentally investigated NOx emissions
rate when hydrogen was burned in a hydrocarbon gas burner, which is used in a wide temperature
range. As a result of the experiments, the amount of NOx when burning hydrogen in a nozzle mixed
burner was twice as high as when burning city gas. However, by increasing the flow velocity of
the combustion air, the amount of NOx could be reduced. In addition, by reducing the number of
combustion air nozzles rather than decreasing the diameter of the air nozzles, a larger recirculation
flow could be formed into the furnace, and the amount of NOx could be reduced by up to 51%.
Furthermore, the amount of exhaust gas recirculation was estimated from the reduction rate of NOx,
and the validity was confirmed by the relationship between adiabatic flame temperature and NOx
calculated from the equilibrium calculation by chemical kinetics simulator software.

Keywords: NOx; hydrogen; exhaust gas self-recirculation; burner






Citation: Kikuchi, K.; Hori, T.;
Akamatsu, F. Fundamental Study on 1. Introduction
Hydrogen Low-NOx Combustion
In order to reduce carbon dioxide emissions from combustion, the use of green fuels,
Using Exhaust Gas Self-Recirculation.
such as hydrogen, ammonia, and eFuel generated from renewable energy sources, is under
Processes 2022, 10, 130. https://
consideration. Among them, hydrogen is generated in the coke oven gas produced in steel
doi.org/10.3390/pr10010130
mills and as a byproduct gas in the production of caustic soda (NaOH) and chlorine gas
Academic Editor: Hans-Joachim (Cl2 ), so there are growing expectations for recovering and using them as energy [1,2]. It is
Gehrmann also expected to be used as a so-called energy carrier to transport renewable energy, which
Received: 16 December 2021
is difficult to transport and store, in the form of carbon-free fuels such as hydrogen and
Accepted: 7 January 2022
ammonia. Hydrogen has a higher burning velocity and lower minimum ignition energy
Published: 9 January 2022
than ammonia, and therefore is a fuel with less risk of misfiring or unburned compared to
ammonia [3–6]. On the other hand, hydrogen has a higher flame temperature than ammonia
Publisher’s Note: MDPI stays neutral
and hydrocarbon fuels, so reducing thermal-NOx is an important challenge [7,8]. There are
with regard to jurisdictional claims in
some reports on the development of industrial burners designed specifically for hydrogen
published maps and institutional affil-
combustion to achieve low-NOx combustion in high temperature conditions, such as
iations.
aluminum melting furnace [9–11]. However, when hydrogen obtained as a byproduct gas
is used as fuel, the supply of hydrogen is not stable, so it is necessary to be able to easily
switch to co-firing with hydrocarbon fuels or hydrocarbon fuels only. If a hydrocarbon
Copyright: © 2022 by the authors.
fueled burner can be modified to enable low-NOx hydrogen combustion with a simple
Licensee MDPI, Basel, Switzerland. modification, rather than a burner structure specifically designed for hydrogen fuel, it will
This article is an open access article not only reduce the cost of modifying existing facilities, but also make it possible to easily
distributed under the terms and switch between hydrocarbon and hydrogen fuels. There are many reports for the formation
conditions of the Creative Commons of NOx in hydrogen flame [12–15]. In diffusion combustion under lean conditions at
Attribution (CC BY) license (https:// low pressure, such as in industrial furnaces, the dominant route of NO formation is the
creativecommons.org/licenses/by/ route produced by the extended Zeldovich mechanism, commonly referred to as thermal-
4.0/). NO [16]. Because of the high activation energy, the reaction of this formation route is

Processes 2022, 10, 130. https://doi.org/10.3390/pr10010130 https://www.mdpi.com/journal/processes

Processes 2022, 10, 130 2 of 14

strongly temperature depended: higher flame temperature, higher residence time, and
higher oxygen concentration lead to higher NO emission [17] (p. 200).
One method of low-NOx combustion in hydrocarbon fueled burners is the use of
exhaust gas recirculation. By circulating the flue gas in the furnace into the combustion air
and reducing the oxygen concentration, the flame temperature is lowered and low-NOx
combustion is achieved [18]. There are two types of exhaust gas recirculation: the one is
the method that uses power such as a blower to circulate the exhaust gas, and the other is
exhaust gas self-recirculation, which uses the momentum of the fuel gas and combustion
air to circulate the exhaust gas. The latter is often used in industrial furnaces. Exhaust gas
recirculation has a trade-off with combustion stability, and if the recirculation rate is set
to an excessive amount, such as more than 20%, it can affect the combustibility and cause
pulsating combustion or misfire due to blowout [17] (p. 202).
Although there have been reports on the effects of hydrogen combustion in a premixed
burner [19–22] and on the effects of exhaust gas recirculation in hydrogen engines [23–25],
there have been no studies on the NO formation characteristics of non-premixed hydrogen
combustion in an industrial furnace and the effects of exhaust gas recirculation have not
yet been studied. Since hydrogen has a very high laminar burning velocity, abnormal
combustion such as knocking in gas engines and backfiring in premixed burners such as
boilers is a problem [24]. On the other hand, burners in industrial furnaces, where most
of the combustion is diffusion at low pressure, do not have such problems. Rather, the
large burning velocity and wide flammability range may not cause abnormal combustion
even if the EGR rate is set higher than that of hydrocarbon fuels. While exhaust gas self-
recirculation is a very useful method for reducing thermal NOx in industrial burners, there
is no systematic report on the effect of the nozzle design on the combustion characteristics
of exhaust gas self-recirculation. In this study, we experimentally evaluated the rate of
increase of NOx when hydrogen was burned in an experimental furnace using the practical
burner with simple structure for hydrocarbon fuels, and the effects of changing the diameter
and number of combustion air nozzles in order to reduce NOx with simple modifications.

2. Experiments and Methods

The configurations of the experimental equipment used in this study are shown in
Figures 1 and 2. The inside dimensions of the experimental furnace are W1200 mm ×
H1200 mm × L3500 mm. The inside of the furnace is surrounded by 260 mm thick ceramic
fiber insulation made of alumina and silica (ISOWOOL BSSR1300, Isolite Industries, Osaka,
Japan). Three ports for R-thermocouples are installed in the upper part of the furnace, and
the ambient temperature at a point 100 mm from the ceiling of the furnace is constantly
measured every five seconds. The axial positions of the thermocouples are Z = 250 mm,
1450 mm, 2600 mm, with Z = 0 mm at the furnace wall on the burner side. The thermocouple
has a strand diameter of 0.5 mm and is covered with a 15 mm diameter ceramic protection
tube. There are 12 tubes for loading at a height of 100 mm above the furnace floor, and each
tube is equipped with equal cooling air to control the temperature at the point Z = 2600 mm
to the required temperature. There is a sight glass on the back wall of the furnace, directly
in front of the burner, so that the flame can be observed. There is an exhaust gas sampling
hole at the side wall of the furnace downstream, which is connected to an exhaust gas
analyzer (PG-340, HORIBA, Kyoto, Japan). The measurement principles were the magnetic
dumbbell method for O2 , non-dispersive infrared absorption method for CO, and the
chemiluminescence method for NOx, respectively.
 composition of city gas 13A is CH4 = 89.6%, C2H6 = 5.62%, C3H8 = 3.40%, and C4H10 = 1.35%
[27].The
[27]. Thelower
lowerheating
heatingvaluevalueofofcity
citygas
gas13A
13Aisiscalculated
calculatedtotobebe40.63
40.63MJ/Nm
MJ/Nm , andthat
3,3and thatofof
HH2 2isis10.83
10.83MJ/Nm
MJ/Nm. City
3 3 . Citygas
gasisissupplied
suppliedfrom
froma apipeline,
pipeline,HH froma acylinder.
2 2from cylinder.After
Afterbeing
being
depressurizedby
depressurized bya aregulator,
regulator,thethemixture
mixtureisisadjusted
adjustedtotothe
thedesired
desiredratio
ratioby
bymass
massflow
flow
controllersand
controllers andsupplied
suppliedatatroomroomtemperature.
temperature.Combustion
Combustionair airisissupplied
suppliedfrom
froma ablower,
blower,
Processes 2022, 10, 130 andthe
and theflow
flowrate
rateisismanually
manuallyadjusted
adjustedby bybutterfly
butterflyvalve
valvetotorequired
requiredflowflowrate
ratevalue 3on
value ofon
14
theturbine
the turbineflow
flowmeter.
meter.

Figure1.1.1.
Figure
Figure Test
Test
Test furnace
furnace
furnace size.
size.
size.

Burner
Burner
HH 2 cylinder
2 cylinder MFC
MFC

13Apipeline
13A pipeline MFC
MFC

Nonreturn
Non returnvalve
valve
Butterflyvalve
Butterfly valve

Turbine
Turbine
Blower
Blower flowmeter
flow meter

Figure
Figure 2.2.2.
Figure Pipe
Pipe flow
flow
Pipe ofof
flow test
test
of furnace
furnace
test burner.
burner.
furnace burner.

We used a diffusion combustion burner of the recirculation flow retention type in this
experiments. As shown in Figure 3, this is a non-premixed burner with eight air nozzles
on a plate at the end of the fuel supply pipe, which forms a stable flame by inducing part
of the fuel gas into the self-recirculating flow formed by the combustion air jetted at high
Processes 2021, 9, x FOR PEER REVIEW 4 of 14
speed. It is one of the most commonly used burners in industrial furnace applications and
is also widely used in actual industrial furnaces [26]. In this study, we adjust combustion
air velocity by replacing or plugging the air nozzles.

Theburner
Figure3.3.The
Figure burnerstructure.
structure.

In the experiment, we measured CO under the low temperature condition of 723 K

and NOx under the high temperature condition of 1273 K at the point of Z = 2600 mm.
Under both conditions, the pressure in the furnace P was adjusted to +15 Pa by the adjust-
ment damper in the exhaust flue and kept constant. The experimental parameters were
the ratio of H2 in the lower heating value of the fuel mixture “EH2” and excess air ratio “λ”.
EH2 is given by Equation (1), where the mole fraction of hydrogen in the fuel mixture is
Processes 2022, 10, 130 4 of 14

In this experiment, we used a mixture of city gas 13A and H2 as fuel. The typ-
ical composition of city gas 13A is CH4 = 89.6%, C2 H6 = 5.62%, C3 H8 = 3.40%, and
C4 H10 = 1.35% [27]. The lower heating value of city gas 13A is calculated to be 40.63 MJ/Nm3 ,
and that of H2 is 10.83 MJ/Nm3 . City gas is supplied from a pipeline, H2 from a cylinder.
After being depressurized by a regulator, the mixture is adjusted to the desired ratio by
mass flow controllers and supplied at room temperature. Combustion air is supplied from
a blower, and the flow rate is manually adjusted by butterfly valve to required flow rate
value on the turbine flow meter.
In the experiment, we measured CO under the low temperature condition of 723 K and
NOx under the high temperature condition of 1273 K at the point of Z = 2600 mm. Under
both conditions, the pressure in the furnace P was adjusted to +15 Pa by the adjustment
damper in the exhaust flue and kept constant. The experimental parameters were the ratio
of H2 in the lower heating value of the fuel mixture “EH2 ” and excess air ratio “λ”. EH2 is
given by Equation (1), where the mole fraction of hydrogen in the fuel mixture is XH2 and
the lower heating values of city gas 13A and hydrogen are LHV 13A and LHV H2 , respectively.

X H2 × LHVH2 × 100
EH2 % = (1)
X H2 × LHVH2 + (1 − X H2 ) × LHV13A

The air nozzles conditions for the experiments are shown in Table 1. V a and V f in the
table show the air flow velocity and fuel gas flow velocity at an excess air ratio of 1.2 during
hydrogen combustion, respectively. ϕ11.9 × 8 is a city gas model condition. In ϕ10.2 × 8
and ϕ8.4 × 8, the air nozzles were replaced with smaller diameter ones to increase the
combustion air velocity by 1.4 and 2 times, respectively. In ϕ11.9 × 4, the air flow velocity
is doubled by reducing the number of air nozzles by half instead of the air nozzle diameter.

Table 1. Air nozzle conditions.

Air Nozzles Number of Air

Gas Velocity Ratio
No. Case Diameter Nozzles
V a /V f
Da [mm] N
1 ϕ11.9 × 8 11.9 8 2.36
2 ϕ10.2 × 8 10.2 8 3.21
3 ϕ8.4 × 8 8.4 8 4.74
4 ϕ11.9 × 4 11.9 4 4.72

3. Results and Discussion

3.1. Combustion Stability
Since hydrogen has a very fast burning velocity and a very low minimum ignition en-
ergy, in non-premixed combustion, it rarely causes abnormal combustion, such as pulsating
combustion, loss of fire, or generation of unburned content [3,5]. Therefore, we evaluated
the combustion stability of each condition based on the amount of CO emitted at 723 K in
city gas combustion (EH2 = 0%).
Figure 4 shows the upper limit of the excess air ratio “λ” at which the burner can burn
at less than 100 ppm CO at EH2 = 0% and 723 K under each condition. In ϕ11.9 × 8 and
ϕ10.2 × 8, combustion was able to continue without CO emission up to the condition of
λ = 1.5 for any combustion rate. However, in ϕ8.4 × 8, CO was emitted over 100 ppm at
λ > 1.15 and 174 kW of input. The amount of CO increased in ϕ11.9 × 4, and the CO was
less than 100 ppm only when λ was smaller than 1.05 with a small combustion rate. In
addition, at 174 kW, misfires occurred and the combustion was very unstable.
 λ =φ10.2
ofand 1.5 for×any
8, combustion
combustionwas able
rate. to continue
However, without
in φ8.4 × 8, CO
COemission up toover
was emitted the 100
condition
ppm
of λ = 1.5 for any combustion rate. However, in φ8.4
at λ > 1.15 and 174 kW of input. The amount of CO increased in φ11.9 × 4, and the ppm
× 8, CO was emitted over 100 CO
at λless
was > 1.15 and
than 100174
ppmkWonly
of input.
whenThe amount
λ was of than
smaller CO increased
1.05 with in φ11.9combustion
a small × 4, and the CO
rate.
was
In less than
addition, 100 kW,
at 174 ppmmisfires
only when λ wasand
occurred smaller than 1.05 with
the combustion wasavery
smallunstable.
combustion rate.
Processes 2022, 10, 130 On the other
In addition, at 174hand, in hydrogen
kW, misfires combustion
occurred (EH2 = 100%was
and the combustion ), there
verywas no pulsating
unstable. 5 of 14
On theorother
combustion hand,
misfire, andinthe
hydrogen combustion
combustion was stable (Eand 100%), thereinwas
H2 = continuous no pulsating
all cases.
combustion or misfire, and the combustion was stable and continuous in all cases.
2.0
2.0
φ 11.9 × 8
11.9××88
φφ10.2
1.8 10.2
1.8
φφ8.4 × 8× 8

[ - [] - ]
8.4 ×× 84
φφ11.9
1.6 φ 11.9 × 4
ratio
ratio
1.6
1.4
airair
1.4
Excess
Excess

1.2
1.2
1.0
50 1.0 0
100 150 200
0
50 Input100 [kW] 150 200
Input [kW]
Figure 4. Combustion stability range (CO < 100 ppm) for different combustion rates under each air
Figure 4. Combustion stability range (CO < 100 ppm) for different combustion rates under each air
nozzle
Figurecase, at 723 K, P stability
4. Combustion = +15 Pa and EH2 = 0%.
nozzle case, at 723 K, P = +15 Parange
and E(CO < 100 ppm) for different combustion rates under each air
H2 = 0%.
nozzle case, at 723 K, P = +15 Pa and EH2 = 0%.
Figure
On the5other
showshand,
the amount of COcombustion
in hydrogen at 723 K and (EλH2= =1.2 underthere
100%), each was
case.no
This figure
pulsating
also shows that a large amount of CO was emitted in ϕ11.9 × 4. It can also
combustion or misfire, and the combustion was stable and continuous in all cases. figure
Figure 5 shows the amount of CO at 723 K and λ = 1.2 under each be
case. seen
This that
also Figure
there shows 5that
was almost anolarge
shows the amount
amount
difference in of
theCO was
COat emitted
723 K and
emission in
λ =ϕ11.9 × 4.between
1.2 under
characteristics Iteach φ11.9
can case.
also be seen
This that
× 8figure
and
also
there shows
was
φ10.2 × 8. that
almost a large
no amount
difference of
in CO
the was
CO emitted
emission in ϕ11.9 ×
characteristics 4. It can
between φ11.9
also be seen
× 8 that
and
there was
φ10.2 × 8. almost no difference in the CO emission characteristics between ϕ11.9 × 8 and
ϕ10.2 × 8.
250
φ 11.9 × 8
250
11.9××88
φφ10.2
200 10.2
φφ8.4 × 8× 8
200
8.4 ×× 84
φφ11.9
] ]

150 11.9 × 4
[ ppm

φ
150
[ ppm

100
COCO

100
50
50
0
0 0 50 100 150 200
0 50 Input100
[kW] 150 200
Input [kW]
Figure 5. CO emissions characteristics for different combustion rates under each air nozzle case. At
723 K, P = +15 Pa, λ = 1.2 and EH2 = 0%.

Figure 6 shows photographs of the flame at EH2 = 0% and 1273 K. In general, in

combustion including hydrocarbons, such as city gas, it is known that the blue–green color
appears in rich-fuel combustion due to the strong band spectrum around 0.51 µm emitted
by C2 radicals, and the blue color appears in lean-fuel combustion due to the strong band
spectrum around 0.39 µm and 0.43 µm emitted by CH radicals [28]. In ϕ11.9 × 8, the entire
inside of the burner tile had a light blue flame, which was characteristic of lean combustion,
while in ϕ8.4 × 8 and ϕ11.9 × 4, the flame was divided into two parts: one part had a clear
dark blue flame that could be visually confirmed and the other part had no visible flame. In
ϕ8.4 × 8 and ϕ11.9 × 4, the flames were unstable, constantly changing the location where
the visual flame could be seen.
 spectrum around
spectrum around 0.39
0.39 µm
µm andand 0.43
0.43 µm
µm emitted
emitted by by CH
CH radicals
radicals [28].
[28]. In φ11.9 ×× 8,
In φ11.9 8, the
the
entire inside of the burner tile had a light blue flame, which was characteristic
entire inside of the burner tile had a light blue flame, which was characteristic of lean of lean
combustion, while
combustion, while in φ8.4 ×× 88 and
in φ8.4 φ11.9 ×× 4,
and φ11.9 4, the
the flame
flame was
was divided
divided into
into two
two parts:
parts: one
one
part had
part had aa clear
clear dark
dark blue
blue flame
flame that
that could
could be
be visually
visually confirmed
confirmed and
and the
the other
other part
part had
had
no visible flame. In φ 8.4 × 8 and φ 11.9 × 4, the flames were unstable,
no visible flame. In φ 8.4 × 8 and φ 11.9 × 4, the flames were unstable, constantly constantly
Processes 2022, 10, 130 6 of 14
changing the
changing the location
location where
where the
the visual
visual flame
flame could
could bebe seen.
seen.

(a)
(a) (b)
(b) (c)
(c) (d)
(d)
Figure 6.
Figure 6. Flame
Flame photographs
photographs at EEH2 == 0%
0% and
and 1273
1273 K
K of
of (a) φ11.9 ×× 8;
(a) φ11.9 8; (b) φ10.2 ×× 8;
(b) φ10.2 8; (c) φ8.4 ×× 8;
(c) φ8.4 8;
Figure 6. Flame photographs atat EH2H2 = 0% and 1273 K of (a) ϕ11.9 × 8; (b) ϕ10.2 × 8; (c) ϕ8.4 × 8;
(d) φ11.9
φ11.9
(d)ϕ11.9 × 4. Input = 174 kW, λ = 1.2, P = +15 Pa.
(d) ××4.4.Input
Input== 174
174 kW,
kW, λ
λ == 1.2,
1.2,PP==+15
+15Pa.
Pa.

Figure777shows
Figure
Figure shows photographs
shows photographs of
photographs of the
of the flame
the flame atatE
flameat EEH2 = 100%
H2 =
H2 =
100%
100% and
andand1273
1273
1273K. K.
K. In In
In hydrogen
hydrogen
hydrogen
combustion,
combustion,the
combustion, the C 2 and CH
theCC22 and CH radicals
CH radicals mentioned
radicalsmentioned above
mentionedabove do
abovedodo not
not
notexist,
exist, and
and
exist, therefore
therefore
and thethe
the
therefore flame
flame
flame
could not
couldnot
could be
notbe visually
bevisually confirmed.
visually confirmed.
confirmed.

(a)
(a) (b)
(b) (c)
(c) (d)
(d)
Figure7.7.
Figure 7.Flame
Flamephotographs
Flame photographsat
photographs atEE
at EH2
H2 = 100% and 1273 KKof
of (a) φ11.9
(a) φ11.9 ×× 8;
ϕ11.9 × 8;8;(b)
(b) Φφ10.2
Φϕ10.2×××8;
8;8;(c)
(c) φ8.4
ϕ8.4××× 8;
(c)φ8.4
Figure H2 == 100%
100%and and1273
1273K of(a) (b)Φφ10.2
8;
(d) (d) φ11.9
8; ϕ11.9 × 4. Input
× 4.×Input
(d) φ11.9 =
= 174
4. Input 174
kW,
= 174 kW,
λ=
kW, λ = 1.2, P = +15
= 1.2,P P= =+15
λ 1.2, Pa.
+15Pa.
Pa.

3.2.
3.2.Temperature
TemperatureDistribution
Distribution
3.2. Temperature Distribution
Figures 8 and 9 show the temperature distribution in the furnace when the temperature
at the point Z = 2600 mm is 1273 K at EH2 = 0% and 100%, respectively. From the figures, it
can be seen that the temperature tends to increase from the front to the back of the furnace
in both EH2 conditions. Comparing EH2 = 0% and 100%, there was no significant difference
in temperature distribution for the same air nozzles condition. Comparing each air nozzles
condition, ϕ11.9 × 8 condition had the lowest temperature upstream of the furnace in both
EH2 conditions.
 Figures at
temperature 8 the
andpoint
9 show the mm
Z = 2600 temperature distribution
is 1273 K at EH2 = 0% andin100%,
the furnace whenFrom
respectively. the
temperature at the point Z = 2600 mm is 1273 K at EH2 = 0% and 100%, respectively. From
the figures, it can be seen that the temperature tends to increase from the front to the back
the figures,
of the it can
furnace inbeboth
seenEthat the temperature tends to increase from the front to the back
H2 conditions. Comparing EH2 = 0% and 100%, there was no
of the furnace in both EH2 conditions. Comparing EH2 = 0% and 100%, there was no
significant difference in temperature distribution for the same air nozzles condition.
Processes 2022, 10, 130
significant
Comparingdifference in temperature
each air nozzles condition,distribution for the same
φ11.9 × 8 condition air lowest
had the nozzlestemperature
condition.
7 of 14
Comparing each air nozzles condition, φ11.9 × 8 condition had the lowest temperature
upstream of the furnace in both EH2 conditions.
upstream of the furnace in both EH2 conditions.
1300
1300
1280

[K]
1280

[K]
Temperature
1260
Temperature 1260
1240 φ 11.9 × 8
1240 φ 11.9 × 8
φ 10.2 × 8
1220 φ 10.2 × 8
φ 8.4 × 8
1220 φ 8.4 × 8
φ 11.9 × 4
1200 φ 11.9 × 4
1200 0 500 1000 1500 2000 2500 3000
0 500 1000Z1500
[mm]2000 2500 3000
Z [mm]
Figure 8.
Figure 8. Temperature
Temperature distribution
distributionwhen
whenthe
thetemperature
temperatureatatthe point
the Z =Z2600
point mmmm
= 2600 is 1273 K atKEat
is 1273 H2
Figure
= 0%. 8. Temperature
Input = 174 kW, distribution
λ = 1.2, P = when
+15 Pa. the temperature at the point Z = 2600 mm is 1273 K at EH2
EH2 = 0%. Input = 174 kW, λ = 1.2, P = +15 Pa.
= 0%. Input = 174 kW, λ = 1.2, P = +15 Pa.
1300
1300
1280
[K]

1280
[K]
Temperature

1260
Temperature

1260
1240 φ 11.9 × 8
1240 φ 11.9 × 8
φ 10.2 × 8
1220 φ 10.2 × 8
φ 8.4 × 8
1220 φ 8.4 × 8
φ 11.9 × 4
1200 φ 11.9 × 4
1200 0 500 1000 1500 2000 2500 3000
0 500 1000Z1500
[mm]2000 2500 3000
Z [mm]
Figure 9. Temperature distribution when the temperature at the point Z = 2600 mm is 1273 K and
Figure
E 9. Temperature
H2 = 100%.
Figure 9. distribution
kW, λ = 1.2, when
Input = 174distribution
Temperature P = +15the
when Pa.temperature
the temperature at
at the point Z
the point Z== 2600
2600 mm
mm is
is 1273
1273 K and
K and
EH2 = 100%. Input = 174 kW, λ = 1.2, P = +15 Pa.
EH2 = 100%. Input = 174 kW, λ = 1.2, P = +15 Pa.
3.3. NOx Emissions Characteristics
3.3. NOx Emissions Characteristics
Figure 10 shows the relationship between NOx and EH2 obtained in a preliminary
Figure
experiment 10 shows the
in another relationship
small test furnacebetween
at 1273 NOx and H2 obtained
and EEH2
K. Although obtained
the burnerin astructure
preliminary
was
experiment in another small test furnace at 1273 Although the
different, the characteristics of a rapid increase in NOx was obtained above EH2 = 80%, was
K. burner structure and
different,
therefore,the
different, wecharacteristics
the characteristics
measured NOx of
ofbyaa rapid
rapid increase
EH2 inin
increase
varying inNOx
5%NOx was
wasobtained
intervalsobtained aboveEEH2
above Eabove
H2 = 80%H2 in=this
80%, and
study.
therefore, we
therefore, we measured
measured NOx
NOx by varyingEEH2
by varying H2 in 5% intervals above EH2H2 ==80%
80%in inthis
thisstudy.
study.
Processes 2021, 9, x FOR PEER REVIEW 8 of 14

Processes 2021, 9, x FOR PEER REVIEW 8 of 14

Processes 2022, 10, 130 8 of 14

100
90
100
80
90
70
80
60

[ppm]
70
50
60
[ppm]
NOxNOx 40
50
30
40
20
30
10
20
0
10
0 20 40 60 80 100
0
0 20 40EH2 [%]60
80 100
E
Figure 10. NOx in a small test [%]
H2 furnace at 1273 K and P = +15 Pa. Input = 11 kW, λ = 1.2.

Figure
Figure10.
10.NOx
NOxininaasmall
smalltest
testfurnace
furnaceatat1273
1273KKand
andPP==+15
+15Pa.
Pa.Input
Input= =1111kW,
kW,λλ= =1.2.
1.2.
Figure 11 shows the amount of NOx for different EH2 under each case. As shown in
Figure 11 in11
Figure
Figure φ11.9
11 shows
shows × the
8,the
NOx gradually
amount
amount of NOx
of NOx increased fromEE47.4
for different
for different ppm each
under
H2under
H2
in
each city gasAs
case.
case. combustion
As shownin
shown in
(E
Figure= 0%)
11 to
in 97.4
ϕ11.9 ×
ppm 8,inNOxhydrogen
gradually combustion
increased (E from= 100%).
47.4
Figure 11 in φ11.9 × 8, NOx gradually increased from 47.4 ppm in city gas combustion
H2 H2 The
ppm reason
in city for
gas this is
combustion that
the
(E adiabatic
(EH2H2 0%)totoflame
==0%) 97.4ppm
97.4 temperature
ppm ininhydrogen
hydrogenof hydrogen
combustion
combustion (2382 K)
(EH2
(E H2== is100%).
aboutThe
100%). 150 Kreason
Thereasonhigher than
for
for thisthat
this of
isisthat
that
the
city adiabatic flame
gas (2233flame
the adiabatic K) [17] (p. 28). In φ10.2
temperature
temperature of hydrogen (2382
× 8, NOx
of hydrogen K) is
wasK)higher
(2382 about
is about 150
than inKφ11.9
K
150 higher than
higher× than that
8 for thatof city
all EofH2
gas (2233 K) [17]
φ8.4 (p. 28). In ϕ10.2 × 8, NOx was higher
φ11.9 than in ϕ11.9 × 8 for all E H2 ranges.
ranges.
city gas In (2233 K)×[17]
8, NOx wasIn
(p. 28). φ10.2
higher than
× 8,inNOx was× higher
8 whenthan in φ11.9
EH2 was small, ×and8 forreversed
all EH2
In ϕ8.4 × 8, NOx was higher than in ϕ11.9 × 8 when EH2 was small, and reversed when
when
ranges.
EH2 was EH2 φ8.4
was
Inlarge, large, resulting
× 8, NOx
resulting inwas in lower
higher
lower NOx. NOx.
than
In in In
ϕ11.9 φ11.9
φ11.9 8×when
× 4,×NOx 4, was
NOxEH2was
waslower
lower small,
than than
in and
ϕ11.9in φ11.9
reversed
× 8 for
×when
8 for
all EH2Eall EH2 ranges.
was
ranges.
H2 large, In φ11.9
resulting
In ϕ11.9 in×lower
× 4, we 4, we NOx.
succeeded
succeeded φ11.9
in reducing
In reducing
in × 4,NOxNOx NOx
inwas inlower
hydrogen hydrogen
than in
combustion φ11.9
combus- to
the
×tion
8 forsame
to all level
theEsame as that
level In
H2 ranges.
in city
φ11.9
as that gas combustion
in ×city gassucceeded
4, we in
combustion ϕ11.9
in in × 8.
φ11.9 ×NOx
reducing 8. in hydrogen combus-
tion to the same level as that in city gas combustion in φ11.9 × 8.
120
φ 11.9 × 8
120
100 φ
φ 10.2 × 8
11.9
100 φ
φ 8.4
10.2××88
80 φ
φ 11.9××84
8.4
[ppm]

80 φ 11.9 × 4
60
[ppm]
NOx

60
40
NOx

40
20
20
0
0 0 40 20 60 80 100
EH2 [%]
0 40 20 60 80 100
Figure 11. NOx for E
different E [%]
under each air nozzle
Figure 11. NOx for differentH2E under each air nozzle
H2 case, at 1273
case, K. Input
at 1273 = 174
K. Input = kW, λ = 1.2,
174 kW, λ =P1.2,
=
H2
+15 Pa.
P = +1511.
Figure Pa.NOx for different EH2 under each air nozzle case, at 1273 K. Input = 174 kW, λ = 1.2, P =
+15 Pa.
Figure
Figure12 12 shows
shows the comparative
comparativeNOxNOxemission
emissionrates
ratesfor
forCases
CasesB–D,
B–D, with
with φ11.9
ϕ11.9 × 8×be-
8
ing 1.0.
being Comparing
1.0.
Figure 12 showscity
Comparing gas
city
the combustion
gas combustion
comparative with
NOx hydrogen
with combustion,
hydrogen
emission rates for CasesitB–D,
combustion, can be
it canseen that that
beφ11.9
with seen NOx
×8
was reduced
NOx was by 20% in ϕ8.4 × 8
in φ8.4 and by 50% in ϕ11.9 ×
in φ11.94. However, in Φϕ10.2 × 8 and
being 1.0. reduced
Comparingby 20%
city gas × 8 and with
combustion by 50%
hydrogen × 4. However,
combustion, inbe
it can Φφ10.2 ×8
seen that
ϕ8.4 × 8, NOx increased by about 20% more than in ϕ11.9 × 8 during city gas combustion.
NOxφ8.4
and × 8, NOxby
was reduced 20% in φ8.4
increased by about 20%by
× 8 and more
50%than in φ11.9
in φ11.9 × 8 duringin
× 4. However, city gas com-
Φφ10.2 ×8
bustion.
and φ8.4 × 8, NOx increased by about 20% more than in φ11.9 × 8 during city gas com-
bustion.
Processes 2021, 9, x FOR PEER REVIEW 9 of 14
Processes 2021, 9, x FOR PEER REVIEW 9 of 14
Processes 2022, 10, 130 9 of 14

1.4
1.4
1.2
1.2
1

A A
raterate
Case
1
0.8

Case
emissions
with
0.8
emissions 0.6
compared with
compared 0.6 φ 11.9 × 8
0.4
NOx

φφ11.9
10.2 ×
× 88
0.4
NOx

0.2 φ 8.4 × 88
φ 10.2 ×
0.2 φ 8.4
φ
11.9××8 4
0 φ 11.9 × 4
0 0 20 40 60 80 100
0 20 EH2 [%]
40 60 80 100
E [%]
Figure 12. NOx emissions rateH2compared with φ11.9 × 8 for different EH2 under each air nozzles
Figure 12. NOx
case at 12.
Figure 1273NOx emissions
K. Input rate compared
= 174 rate
emissions kW, compared
λ with
= 1.2, P =with φ11.9×× 88 for
+15 Pa.
ϕ11.9 fordifferent
differentEEH2 under each
H2 under each air
air nozzles
nozzles
case at
case at 1273
1273 K.
K. Input
Input == 174 kW,
kW, λλ == 1.2, P = +15
+15 Pa.
Pa.
3.4. Relationship between Excess Air Ratio “λ” and NOx
3.4.
3.4. Relationship
Relationship between
between Excess
Excess Air
Air Ratio “λ”
“λ” and
Ratiobetweenand NOx
NOx
Figure 13 shows the relationship λ and NOx in city gas combustion (EH2 =
Figure
0%). Figure 13
In city 13 shows
gasshows the relationship
the relationship
combustion, between
NOx increasedbetween λ and
or λ NOx
and
remained NOxinalmost
city gasflat
in city combustion
gasascombustion(EH2 =
λ increased. 0%).
(EThis
H2 =
In
0%).city gas
In city
tendency combustion,
gas combustion,
is consistent NOx
with the increased
NOx or
increased
generally remained
knownor remainedalmost flat as
almost flatof
NOx characteristics λ increased.
ashydrocarbon This
λ increased. fuelsten-
This
dency
[17] (p.is209).
tendency consistent
Figurewith
is consistent the the
14with
shows generally known
generally
the known
relationship NOx characteristics
NOx
between characteristics
λ and NOx of hydrocarbon
ofhydrogen
in hydrocarbon fuels [17]
fuels
combus-
(p.
[17] 209). Figure 14 shows the relationship between λ and NOx in hydrogen combustion
tion (p.
(EH2209). Figure
= 100%). In14allshows
cases, the
thererelationship betweenchange
was no significant λ and NOx in theinNOxhydrogen
when combus-
λ was in
(EH2 = 100%). In all cases, there was no significant change in the NOx when λ was in
tion (E
the range H2 = 100%). In all cases, there was no significant change in the NOx when λ was in
range of φ
the of 1.05
1.05to
to1.2.
1.2.On Onthetheother hand,
other when
hand, whenλ wasλ wasset to
set1.4toor1.4
higher, the NOx
or higher, the inNOx
the
11.9
in range
× 8 and
ϕ11.9 of 1.05 to
φ10.2
× 8 and 1.2.
ϕ10.2 On the
× 8 tended
× 8 tended other hand,
to decrease. when λ
This can
to decrease. was set
Thisbecansaidto 1.4
betosaid or higher,
be atospecific the NOx in
characteristic
be a specific charac-φ
11.9 × 8of
teristic
of hydrogen and φ10.2
hydrogen ×combustion,
combustion,8 tended
which to was
decrease.
whichnotwas This
seennotincan
citybegas
seen insaid
city togas
be combustion.
combustion. a specific characteristic
Generally, Generally,
in pre-
in
of premixed
hydrogen combustion
combustion, of hydrocarbon
which was not fuels,
seen inNOx
city peaks
gas around
combustion.
mixed combustion of hydrocarbon fuels, NOx peaks around λ of 1.0, and NOx decreases λ of 1.0, and
Generally, NOx
in de-
pre-
creases
mixed as
as λ increases λ increases
combustion [17,29]
of(pp.
[17,29] (pp.
hydrocarbon 207–209). Since
fuels, hydrogen
207–209). Since NOx peaks hydrogen
around
has has
very fast very
λ ofburningfast burning
1.0, andvelocity, velocity,
NOx decreases
its NOx
its
as λNOx
emissions emissions
increases characteristics
[17,29] (pp. in
characteristics 207–209).
diffusioninSince
diffusion combustion
hydrogen
combustion also may
has very
may alsotobethose
fastclose
be burning close oftopremixed
velocity, those
its NOx of
premixed
emissions
combustion. combustion.
characteristics in diffusion combustion may also be close to those of premixed
combustion.
100
100
80
[ppm]

80
[ppm]

60
O₂)O₂)

60
0%

40
0%

φ 11.9 × 8
40
(@(@

φ 11.9
φ 10.2 ×
× 88
20
NOx

φ 8.4 × 88
φ 10.2 ×
20
NOx

φ 8.4 × 8
φ 11.9 × 4
0 φ 11.9 × 4
0 1.0 1.1 1.2 1.3 1.4 1.5 1.6
1.0 1.1 Excess
1.2 air 1.4λ [-]1.5
1.3ratio 1.6
Excess air ratio λ [-]
Figure 13.
Figure 13. NOx
NOx for
for different
different λλunder
undereach
eachair
airnozzle
nozzlecase,
case,at
at1273
1273KKand
andEEH2 = 0%.
H2 = 0%. Input
Input =
= 174
174 kW,
kW,
P = +15
Figure Pa.
13.
P = +15 Pa. NOx for different λ under each air nozzle case, at 1273 K and E H2 = 0%. Input = 174 kW,

P = +15 Pa.
Processes 2021, 9, x FOR PEER REVIEW 10 of 14
Processes 2022, 10, 130 10 of 14

140
120

NOx (@ 0 % O2) [ppm]

100
80
60
φ 11.9 × 8
40 φ 10.2 × 8

20 φ 8.4 × 8

φ 11.9 × 4
0
1.0 1.1 1.2 1.3 1.4 1.5 1.6
Excess air ratio λ [-]
Figure 14.
Figure 14. NOx
NOx for
for different
differentλλ under
undereach
eachair
airnozzle
nozzlecase,
case,atat1273
1273KKand
andEEH2 = 100%.
H2 = 100%. Input = 174
Input = 174 kW,
kW,
P = +15 Pa.
P = +15 Pa.

3.5.
3.5. Relationship
Relationship between
between Adiabatic
Adiabatic Flame
Flame Temperature
Temperature“T
“Tf ”f” and
and NOx
NOx
Flame
Flame temperature
temperature is is one
one ofof the
the most
most important
important properties
properties to to consider
consider in in thermal
thermal
NOx
NOx emissions
emissions characteristics.
characteristics. Adiabatic flame temperature is often often used
used to to evaluate
evaluate the the
properties
properties of of fuels.
fuels. Figure 15 shows shows the the relationship
relationship between
between NOx NOx andand adiabatic
adiabatic flame
flame
temperature
temperature“T “Tff”,
”, which
which is is strongly
strongly related
related to to the
the formation
formation of Thermal-NO.TTff at
of Thermal-NO. at ambient
ambient
pressure
pressure and and temperature was simulated using CHEMIKIN II with changing EE
simulated using CHEMIKIN II with changing H2H2andandλ.
In exhaust
In exhaust
λ. gasgas self-recirculation,
self-recirculation, it is itdifficult
is difficult to measure
to measure recirculation
recirculation rate, rate,
so that sothese
that
these adiabatic
adiabatic flameflame temperatures
temperatures were weresimulatedsimulated
withoutwithout considering
considering the exhaust
the exhaust gas
gas recir-
recirculation.
culation. As shown As shown in Figure
in Figure 15, NOx
15, NOx rapidly
rapidly increased
increased as T Tf increased
asf increased above
above 21002100 K
K in
in ϕ11.9 × 8 and ϕ10.2 × 8. On the other hand, in ϕ8.4 ×
φ11.9 × 8 and φ10.2 × 8. On the other hand, in φ8.4 × 8 and φ11.9 × 4, NOx increased 8 and ϕ11.9 × 4, NOx increased
linearly with increasing Tf and there is no inflection point. Therefore, in ϕ8.4 × 8 and
linearly with increasing Tf and there is no inflection point. Therefore, in φ8.4 × 8 and φ
ϕ11.9 × 4, the actual flame temperature was considered to be even lower due to exhaust
11.9recirculation.
gas × 4, the actual In flame temperature
addition, the slopes was in considered
the graphs to of
be ϕ8.4
even × lower
8 anddue to exhaust
ϕ11.9 × 4 were gas
recirculation.
almost equal. In Inaddition, the slopes
general, reports oninthe thelow-NOx
graphs ofeffectφ8.4of× exhaust
8 and φ11.9 × 4 were almost
gas recirculation for
hydrocarbon
equal. In general, fuelsreports
show that NOx
on the was reduced
low-NOx effect by about 50%
of exhaust gasand 20% whenfor
recirculation the exhaust
hydrocar-
gas
bonrecirculation
fuels show that rateNOx(EGRwas rate) was setby
reduced to about
10% and 50% 5%,andrespectively
20% when[17] (pp. 202–203).
the exhaust As
gas recir-
shown
culation inrate
Figure
(EGR 12,rate)
in ϕ8.4was×set 8 and ϕ11.9
to 10% and ×5%,4, NOx in EH2 = [17]
respectively 100%(pp.202–203).
was reducedAs byshown
about
20% and 50%,
in Figure 12, inrespectively.
φ8.4 × 8 and Therefore,
φ11.9 ×we 4, assumed
NOx in EEGR rates of
H2 = 100% was5%reduced
and 10%by forabout
ϕ8.4 20%×8
and ϕ11.9 × 4, respectively, and re-plotted the adiabatic flame temperature accounting for
and 50%, respectively. Therefore, we assumed EGR rates of 5% and 10% for φ8.4 × 8 and
exhaust gas recirculation, which is shown in Figure 16. It was found that the curves were
φ11.9 ×all4,the
almost respectively,
same, although and re-plotted
the NOx for the ϕ10.2
adiabatic× 8flame
and ϕ8.4temperature
× 8 wereaccounting
slightly higher for ex-
in
haust gas recirculation, which is shown in Figure 16. It
Tf < 2100 K. This curve shows that the Tf in ϕ11.9 × 4 was below 2100 K due to exhaustwas found that the curves were
gas recirculation
almost all the same, even during the
although hydrogen
NOx for φ10.2 × 8 and
combustion, which φ8.4
suppressed
× 8 were the increase
slightly in
higher
NOx. The reason for the higher NOx in ϕ10.2 × 8 and ϕ8.4
in Tf < 2100 K. This curve shows that the Tf in φ11.9 × 4 was below 2100 K due to exhaust × 8 when T f was small (E H2
was small) was probably due to the lower local combustion air ratio caused by the larger
gas recirculation even during hydrogen combustion, which suppressed the increase in
recirculation flow formed near the air nozzles due to the increase of combustion air velocity,
NOx. The reason for the higher NOx in φ10.2 × 8 and φ8.4 × 8 when Tf was small(EH2
which increased the amount of fuel gas drawn near the air nozzles. The decrease in the
was small)
local combustionwas probably
air ratiodue to the lower
is expected local combustion
to increase the flame air ratio caused
temperature and bylead
the larger
to an
recirculation flow formed near the air nozzles due to the increase
increase in NOx generation. The decrease in the local combustion air ratio is also consistent of combustion air veloc-
ity, which
with the darkerincreased the amount
blue color in the flameof fuel gas drawninnear
photograph Figurethe6.air
Asnozzles. The decrease
EH2 increased, the localin
the local combustion air ratio is expected to increase the flame temperature
air ratio bias decreased as the fuel gas velocity increases, since less fuel gas was drawn into and lead to an
increase
the in NOx generation. The decrease in the local combustion air ratio is also con-
air nozzles.
sistent with the darker blue color in the flame photograph in Figure 6. As EH2 increased,
the local air ratio bias decreased as the fuel gas velocity increases, since less fuel gas was
drawn into the air nozzles.
 Processes
Processes 2021,
2022, 9,
10,x130
FOR PEER REVIEW 1111ofof 14
14
Processes 2021, 9, x FOR PEER REVIEW 11 of 14

120
120 φ 11.9 × 8
φ 11.9 × 8
100 φ 10.2 × 8
100 φ 10.2 × 8
φ 8.4 × 8
φ 8.4 × 8
80 φ 11.9 × 4
80
[ppm]
φ 11.9 × 4

NOx[ppm] 60
60
NOx

40
40
20
20
0
0 1800 1900 2000 2100 2200 2300
1800 1900 2000 2100 2200 2300
T [K]
Tff[K]
Figure
Figure 15.
15. NOx
NOxforfordifferent
differentTfTwithout accounting for EGR at 1273 K. Input = 174 kW. λ = 1.2, P =
f without accounting for EGR at 1273 K. Input = 174 kW. λ = 1.2,
Figure
+15 Pa.15.
E NOx for different Tf without
H2 = 0, 50, 80, 85, 90, 95, 100%.
accounting for EGR at 1273 K. Input = 174 kW. λ = 1.2, P =
P = +15 Pa. EH2 = 0, 50, 80, 85, 90, 95, 100%.
+15 Pa. EH2 = 0, 50, 80, 85, 90, 95, 100%.
120
120
100
100
80
80
[ppm]
NOx[ppm]

60
60
NOx

40
40 φ 11.9 × 8 without accounting for EGR
φφ11.9
10.2××88without
withoutaccounting
accountingfor
forEGR
20 EGR
20 φφ 10.2
8.4 ×× 88with
without accounting
accounting for EGR
for EGR
φφ 8.4 ××
11.9 8 with accounting
4 with forfor
accounting EGR
EGR
0 φ 11.9 × 4 with accounting for EGR
0 1800 1900 2000 2100 2200 2300
1800 1900 2000 2100 2200 2300
T [K]
Tff[K]
Figure 16. NOx for different Tf accounting for EGR at 1273 K. Input = 174 kW. λ = 1.2. P = +15 Pa.
Figure
EFigure 16.
50,NOx
H2 = 0, 16. NOx for
80, 85,for different
90,different TTf faccounting
95, 100%. accountingfor
forEGR
EGRatat1273
1273K.
K.Input
Input==174
174kW.
kW.λλ==1.2.
1.2.PP==+15
+15Pa.
Pa.
EEH2 = =
0,0,
50,50,
80,80,
85,85,90,90,95, 100%.
95, 100%.
H2
Comparing φ8.4 × 8 and φ11.9 × 4, the gas velocity ratio Va/Vf is the same, however,
Comparing φ8.4
Comparing ϕ8.4 × × 88 and φ11.9×
and ϕ11.9 × 4, the gas velocity ratio Va/V /Vf isis
the
thesame,
same,however,
however,
their combustion characteristics were very different. This is duea to fthe different distance
theircombustion
their combustioncharacteristics
characteristicswere wereveryverydifferent.
different.This Thisisisdue
duetotothethedifferent
differentdistance
distance
between
between the
the air
air nozzles
nozzles ininφ8.4
ϕ8.4 ××88and
and φ11.9
ϕ11.9 ××4.4.Figure
Figure17 17
shows
shows nozzle
nozzlelayout of φ
layout of
between the air nozzles in φ8.4 × 8 and φ11.9 × 4. Figure 17 shows nozzle layout of φ
8.4
ϕ8.4× 8×and8 and φ11.9
ϕ11.9 × 4.
× Defining
4. Defining thethedistances
distances between
between the the
air nozzles
air nozzles in φ8.4
in ϕ8.4× 8×and φ
8 and
8.4 × 8 and φ11.9 × 4. Defining the distances between the air nozzles in φ8.4 × 8 and φ
ϕ11.9
11.9 × 4×as4 Las Lc and
c and Ld , respectively,
Ld, respectively, we obtain Ld/Lc L=d1.85.
we obtain /Lc =In1.85. In general,
general, when multiple
when multiple jets are
11.9
jets × 4
are as L c and Ld, respectively, we obtain Ld/Lc = 1.85. In general, when multiple jets are
placed in a forest, the space closed between the jets becomes
placed in a forest, the space closed between the jets becomes depressurized and a recircu- depressurized and
placed in a
a recirculation
lation forest,
flow is formed the
flow isspace
formed
there closed between
[30]. there
Jianchun the jets
[30].et Jianchunbecomes
al. developed depressurized
et al.andeveloped
approximate and a
an equationrecircu-
approximate that
lation flowthat
equation
expresses is formed
the expresses there
recirculation the[30].
rate Jianchun
recirculation
in terms of etvarious
rateal.
indeveloped
termsjet of an approximate
various
parameters, jetincluding equation
parameters, that
theincluding
distance
expresses
the distance
between the between
jets, recirculation
where the jets, rate
where in terms
recirculation of and
various
the recirculation
rate jet
rate parameters,
the distance andbetween including
the distance the distance
between
jets are jets are
proportional
between
proportionaljets, where
[31]. the recirculation
Therefore, in rate
ϕ11.9 ×and
4, the
the distance
EGR between
rate was jets
higher, are
and proportional
the oxygen
[31]. Therefore, in φ11.9 × 4, the EGR rate was higher, and the oxygen concentration in
[31]. Therefore, in
concentration φ11.9
in the × 4, the EGR
combustion ratelower,
air was was higher,
resulting andinthe oxygencombustion
unstable concentration in
to the
the combustion air was lower, resulting in unstable combustion to the level that CO was
level
the that
combustion CO was emitted at low temperatures, and it is thought that NOx was reduced by
emitted at low air was lower, resulting
temperatures, in unstable
and it is thought thatcombustion to the level
NOx was reduced by that CO was
lowering the
lowering
emitted the flame temperature
at low temperatures, even in hydrogen
and it iscombustion. combustion.
thought that NOx was reduced by lowering the
flame temperature even in hydrogen
flame temperature even in hydrogen combustion.
Processes 2021, 9, x FOR PEER REVIEW 12 of 14
Processes 2022, 10, 130 12 of 14

(a) (b)
FigureFigure
17. Nozzle layoutlayout
17. Nozzle φ8.4
of (a) of × 8;φ11.9
× 8; (b)
(a) ϕ8.4 × 4. × 4.
(b) ϕ11.9

4. Conclusions
4. Conclusions
We experimentally evaluated the combustion characteristics of exhaust gas self-
We experimentally evaluated the combustion characteristics of exhaust gas self-re-
recirculation combustion and the effect of combustion air nozzle parameters using re-
circulation combustion and the effect of combustion air nozzle parameters using recircu-
circulation flow retention type burner, and examined the possibility of realizing low-NOx
lation flow retention type burner, and examined the possibility of realizing low-NOx hy-
hydrogen combustion. The following conclusions can be obtained.
drogen combustion. The following conclusions can be obtained.
(1) When hydrogen was burned using a burner with city gas specifications, NOx increased
(1) Whentwice hydrogen
as much wasdueburned
to the using a burner
increase in flamewith city gas specifications, NOx in-
temperature.
creased twice as much due to the increase in flame temperature.
(2) When the air nozzles diameter was reduced and the air flow velocity was increased
(2) Whenby the air nozzles
a factor of 1.4,diameter
the amount wasofreduced
fuel gasand drawnthe air
intoflow velocity
the air nozzlewas increased
increased and the
by a factor of 1.4, the amount of fuel gas drawn
local air ratio decreased, resulting in an increase in NOx. into the air nozzle increased and the
local
(3) airWhen ratio decreased,
the combustion resulting in andiameter
air nozzle increasewas in NOx.
reduced and the combustion air flow
(3) Whenvelocity
the combustion air nozzle diameter was
was doubled, NOx increased under low hydrogen reduced and the combustion
mixing ratio airconditions
flow
velocity was doubled, NOx increased under low hydrogen
as same as mentioned above, but NOx reduced under high hydrogen mixing ratio mixing ratio conditions
as same as mentioned
conditions due to above,
exhaustbut gasNOx reduced under high hydrogen mixing ratio
recirculation.
conditions
(4) Whendue to exhaust
the number gasnozzles
of air recirculation.
was reduced to half instead of the air nozzles diameter
(4) Whenand the the
number of air nozzles
combustion was reduced
air velocity was doubled,to halfNOxinstead
was of the air at
reduced nozzles diam-rates,
all co-firing
eter and the combustion air velocity was doubled, NOx was
and a 50% reduction was successfully achieved in hydrogen combustion. And the reduced at all co-firing
rates, NOx
and ain50% thisreduction
case was was almost successfully
equal to the achieved
NOx ininthe hydrogen combustion.
case of city And in
gas combustion
the NOx in this case
the city gas model. was almost equal to the NOx in the case of city gas combustion
in
(5)theBycityreducing
gas model.the number of air nozzles, the distance between air nozzles became
(5) By reducing the number
larger, resulting in a of air nozzles,
larger EGR ratio. theAsdistance
a result,between air nozzles
misfire occurred became
in hydrocarbon
larger,combustion,
resulting inbut a larger EGR ratio.
in hydrogen As a result,stable
combustion, misfire occurred in
combustion hydrocarbon
could be continued
combustion,
withoutbut in hydrogen
abnormal combustion. combustion, stable combustion could be continued
without
(6) When abnormal combustion.
the number of air nozzles was reduced to half, the exhaust gas recirculation rate
(6) Whenwas theabout
number 10%,of which
air nozzles was reduced
was assumed from tothehalf, the exhaust
adiabatic gas recirculation
flame temperature simulated
by the
rate was aboutequilibrium
10%, which calculation
was assumedof chemical
from the kinetics simulator
adiabatic flame software.
temperature sim-
(7) Itby
ulated was theconfirmed
equilibrium thatcalculation
low-NOx combustion
of chemicalby exhaust
kinetics gas recirculation
simulator software.is possible in
(7) It washydrogen
confirmed combustion.
that low-NOx combustion by exhaust gas recirculation is possible
in hydrogen combustion.
Author Contributions: Conceptualization, K.K.; software, K.K.; validation, K.K.; formal analy-
Authorsis,Contributions: Conceptualization,
K.K.; investigation, K.K.;
K.K.; resources, software,
K.K. K.K.;
and F.A.; validation, K.K.;
writing—original formal
draft analysis, K.K.;
preparation,
K.K.; writing—review
investigation, K.K.; resources, K.K. and F.A.; writing—original draft preparation, K.K.; writ-
and editing, T.H. and F.A.; supervision, F.A.; project administration, F.A. All authors
ing—review and editing, T.H. and F.A.; supervision, F.A.; project
have read and agreed to the published version of the manuscript. administration, F.A. All authors
have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Funding: This research received no external funding.
Acknowledgments: We would like to express my appreciation to YOKOI KIKAI KOSAKUSYO Co.,
Ltd. and TAKAMITSU Co., Ltd. for their assistance in the construction of the experimental apparatus.
Conflicts of Interest: The authors declare no conflict of interest.
Processes 2022, 10, 130 13 of 14

References
1. Verhage, A.J.; Coolegem, J.F.; Mulder, M.J.; Yildirim, M.H.; de Bruijn, F.A. 30,000 h operation of a 70 kW stationary PEM fuel cell
system using hydrogen from a chlorine factory. Int. J. Hydrogen Energy 2013, 38, 4714–4724. [CrossRef]
2. Haneda, T.; Akisawa, A. Technological assessment of PEFC power generation system using by-product hydrogen produced from
a caustic soda plant. Int. J. Hydrogen Energy 2017, 42, 3240–3249. [CrossRef]
3. Pareja, J.; Burbano, H.J.; Amell, A.; Carvajal, J. Laminar burning velocities and flame stability analysis of hydrogen/air premixed
flames at low pressure. Int. J. Hydrogen Energy 2011, 36, 6317–6324. [CrossRef]
4. Takizawa, K.; Takahashi, A.; Tokuhashi, K.; Kondo, S.; Sekiya, A. Burning velocity measurements of nitrogen-containing
compounds. J. Hazard. Mater. 2008, 155, 144–152. [CrossRef]
5. Ono, R.; Nifuku, M.; Fujiwara, S.; Horiguchi, S.; Oda, T. Minimum ignition energy of hydrogen–air mixture: Effects of humidity
and spark duration. J. Electrost. 2007, 65, 87–93. [CrossRef]
6. Sadaghiani, M.S.; Arami-Niya, A.; Zhang, D.; Tsuji, T.; Tanaka, Y.; Seiki, Y.; May, E.F. Minimum ignition energies and laminar
burning velocities of ammonia, HFO-1234yf, HFC-32 and their mixtures with carbon dioxide, HFC-125 and HFC-134a. J. Hazard.
Mater. 2020, 407, 124781. [CrossRef] [PubMed]
7. Choudhuri, A. Characteristics of hydrogen–hydrocarbon composite fuel turbulent jet flames. Int. J. Hydrogen Energy 2003, 28,
445–454. [CrossRef]
8. Choudhuri, A.R.; Gollahalli, S.R. Computational and Experimental Study of the Structure of Hydrocarbon-Hydrogen Turbulent
Jet Flame. In Proceedings of the 41st Aerospace Sciences Meeting and Exhibit, Reno, NV, USA, 6–9 January 2003; p. AIAA
2003-335. [CrossRef]
9. Taguchi, S.; Takahashi, T.; Taniyama, K.; Nakagaki, H. Development of hydrogen combustion technology in industrial burners. In
Proceedings of the Symposium (Japanese) on Combustion, Osaka, Japan, 14–16 November 2018; p. C342.
10. Toyota Develops World’s First General-Purpose Hydrogen Burner for Industrial Use. Available online: https://global.toyota/en/
newsroom/corporate/25260001.html (accessed on 25 October 2021).
11. Yamaguchi, M.; Haneji, T.; Nakabayashi, H.; Hagihara, Y. Evaluation of NOx emission from Hydrogen-Oxygen Com-bustion in
the Industrial Furnace. In Proceedings of the Symposium (Japanese) on Combustion, Online, 22–24 November 2021; p. B333.
12. Konnov, A. On the relative importance of different routes forming NO in hydrogen flames. Combust. Flame 2003, 134,
421–424. [CrossRef]
13. Frassoldati, A.; Faravelli, T.; Ranzi, E. A wide range modeling study of NOx formation and nitrogen chemistry in hydrogen
combustion. Int. J. Hydrogen Energy 2006, 31, 2310–2328. [CrossRef]
14. Skottene, M.; Rian, K.E. A study of NOx formation in hydrogen flames. Int. J. Hydrogen Energy 2007, 32, 3572–3585. [CrossRef]
15. Purohit, A.L.; Nalbandyan, A.; Malte, P.C.; Novosselov, I.V. NNH mechanism in low-NOx hydrogen combustion: Experimental
and numerical analysis of formation pathways. Fuel 2021, 292, 120186. [CrossRef]
16. Zeldovich, Y.B. The Oxidation of Nitrogen in Combustion and Explosions. Acta Physicochim. USSR 1946, 21, 577–628.
17. Nakamachi, I.; Takei, M.; Hase, T.; Moriyama, T. Gas Nensho No Riron to Jissai, 2nd ed.; The Energy Conservation Center: Tokyo,
Japan, 2012. (In Japanese)
18. Delabroy, O.; Haile, E.; Lacas, F.; Candel, S.; Pollard, A.; Sobiesiak, A.; Becker, H. Passive and active control of NOx in industrial
burners. Exp. Therm. Fluid Sci. 1998, 16, 64–75. [CrossRef]
19. Van Der Drift, A.; Tjeng, S.; Beckers, G.; Beesteheerde, J. Low-NOx hydrogen burner. Int. J. Hydrogen Energy 1996, 21,
445–449. [CrossRef]
20. Rørtveit, G.J.; Zepter, K.; Skreiberg, Ø.; Fossum, M.; Hustad, J.E. A comparison of low-NOx burners for combustion of methane
and hydrogen mixtures. Proc. Combust. Inst. 2002, 29, 1123–1129. [CrossRef]
21. Heffel, J.W. NOx emission reduction in a hydrogen fueled internal combustion engine at 3000 rpm using exhaust gas recirculation.
Int. J. Hydrogen Energy 2003, 28, 1285–1292. [CrossRef]
22. Wei, Z.; Wang, L.; Liu, H.; Liu, Z.; Zhen, H. Numerical Investigation on the Flame Structure and CO/NO Formations of the
Laminar Premixed Biogas–Hydrogen Impinging Flame in the Wall Vicinity. Energies 2021, 14, 7308. [CrossRef]
23. Rao, A.; Wu, Z.; Mehra, R.K.; Duan, H.; Ma, F. Effect of hydrogen addition on combustion, performance and emission of
stoichiometric compressed natural gas fueled internal combustion engine along with exhaust gas recirculation at low, half and
high load conditions. Fuel 2021, 304, 121358. [CrossRef]
24. Guo, H.; Zhou, S.; Zou, J.; Shreka, M. A Numerical Investigation on De-NOx Technology and Abnormal Combustion Control for
a Hydrogen Engine with EGR System. Processes 2020, 8, 1178. [CrossRef]
25. Xin, Y.; Wang, K.; Zhang, Y.; Zeng, F.; He, X.; Takyi, S.A.; Tontiwachwuthikul, P. Numerical Simulation of Combustion of Natural
Gas Mixed with Hydrogen in Gas Boilers. Energies 2021, 14, 6883. [CrossRef]
26. The Japan Society of Mechanical Engineers. JSME Combustion Handbook; The Japan Society of Mechanical Engineers: Tokyo, Japan,
1995; pp. 158–160.
27. Tokyo Gas Co., Ltd. Heating Value of Tokyo Gas. Available online: https://home.tokyo-gas.co.jp/gas/userguide/shurui.html
(accessed on 21 November 2021).
28. Mizutani, Y. Nensho Kougaku; Morikita Publishing Co., Ltd.: Tokyo, Japan, 1977; p. 194. (In Japanese)
29. Cho, E.-S.; Chung, S.H. Numerical evaluation of NOx mechanisms in methane-air counterflow premixed flames. J. Mech. Sci.
Technol. 2009, 23, 659–666. [CrossRef]
Processes 2022, 10, 130 14 of 14

30. Fujimine, T.; Nakamachi, I.; Hayakawa, A. Combustion Stabilization of Gas Burners. J. Combust. Soc. Jpn. 2009, 51, 85–93.
(In Japanese) [CrossRef]
31. Mi, J.; Li, P.; Zheng, C. Impact of injection conditions on flame characteristics from a parallel multi-jet burner. Energy 2011, 36,
6583–6595. [CrossRef]