Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
Premixed Combustion of Hydrogen and Syngas Fuels in Gas Turbine
Combustors
Onur Tuncer
Istanbul Technical University, Dept. of Aeronautical Engineering, Maslak, 34469, Istanbul, Turkey
Power generation industry often relies on land based gas turbine engines for energy conversion from liquid and/or gaseous
fuels. When it comes to gaseous fuels, from a historical perspective, combustors were often designed for methane fuel,
since it is the major constituent of natural gas, an abundant fuel resource. Nevertheless, with the recent proliferation of the
IGCC (Integrated Gasified Combined Cycle) technology, the use of hydrogen rich fuels is dictated for many power
generation installations. Fuel flexibility is a major issue for these gas turbine engines. For the last two decades lean
premixed combustion has become the industry standard for many applications as it enables single digit NOx emissions.
Hydrogen enrichment significantly alters flame behaviour of hydrocarbon fuels due to its much faster propagation speeds.
This presents challenges in terms of flame holding, flashback and thermo-acoustic instability.
There are several methods which the power generation industry utilizes in order to alleviate this situation. Each has
advantages and disadvantages in terms of key performance metrics such as volumetric heat release, pattern factors and
emissions. This paper aims to present the state-of-the-art on hydrogen rich fuel combustion in gas turbine environment
with examples from the open literature. Firstly, conventional swirl stabilized hydrogen enriched flames are presented with
a thorough discussion on associated challenges; secondly a novel technique that relies on aerodynamic flame stabilization
is discussed. This second technique is also termed as low-swirl combustion. Benefits and possible drawbacks for each
premixed flame configuration are discussed.
Keywords: hydrogen; syngas; combustion; gas turbine engine; flame holding
1. Introduction
Synthesis gas (syngas), an environmentally clean source of energy, is a variable mixture of primarily hydrogen and
carbon monoxide and some other species such as hydrogen sulphur, hydrogen cyanide, hydrogen chloride and ammonia
in smaller proportions. Syngas is either obtained from gasification of coal or biomass. Depending on the gasification
process, and which solid is gasified substantial changes in the resulting syngas composition can occur [1]. Variability of
the syngas composition (Table 1) can significantly alter the flame behaviour [2]. For example, an increase in hydrogen
proportion will assure better flame stability but, at the very same time, it will make the combustor more susceptible to
flashback and thermo-acoustic instability. In addition it will also increase nitric oxide emissions due to higher
temperatures. On the other hand an increase in carbon monoxide content will deteriorate flame stability. Consequently,
it is necessary that a syngas combustor be able to tackle the variability in fuel composition without requiring any design
changes [3].
Low BTU fuel as syngas has less chemical bonding energy per unit weight. Therefore, in order to achieve a desired
power output from syngas high mass flow rates needs to be used [4]. High mass flow rates normally translate into
higher injection speeds, which pose a problem in flame holding. Furthermore, the main reaction zone residence time
determines the pollutant levels [5-6]. For the combustion of syngas high turbulence intensity (between 10-20%) at the
burner exit is needed since flame speed is related to the turbulence level in the main reaction zone. Therefore the design
of the syngas injection nozzle and combustor should incorporate the challenges posed by flame holding, low BTU per
unit mass and emissions.
Table 1 Syngas composition variation across different sites in USA [2].
H2
CO
CH4
CO2
N2+Ar
H 2O
H2/CO Ratio
Diluent
LHV (kJ/m3)
946
PSI
24.8
39.5
1.5
9.3
2.3
22.7
0.63
Steam
8224
Tampa
37.2
46.6
0.1
13.3
2.5
0.3
0.80
N2
9962
Sierra-Pacific
14.5
23.6
1.3
5.6
49.3
5.7
0.79
Steam
5024
©FORMATEX 2013
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
In two separate works Borghi and Peters [7-8] have identified a regime diagram for premixed flames which is shown
in Figure 1. Note that practical combustors mostly operate within the flamelet regime. Here reaction takes place in a
thin front that separates the products and the reactants.
Syngas utilization in a multi fuel combustion environment is another challenge that needs to be addressed. In many
of the existing installations, other fuels like natural gas and/or fuel oil are readily available on site. Therefore it is
desirable to understand how mixing natural gas and syngas in different proportions effects flame holding, flashback,
and combustion dynamics.
For premixed syngas injection, one of the biggest challenges is flashback, since hydrogen flame speeds are quite high
[9-10]. Flashback into the pre-mixer section leads to thermal overload and destruction of the hardware therefore it must
be avoided at all load conditions [11]. Flashback can be prevented by using specially designed flame holders or by
injecting syngas in a separate non-premixed arrangement. However, in transitioning from natural gas as the fuel of
choice to syngas, it is desirable to keep hardware changes to a minimum, given the extensive body of knowledge with
current natural gas related hardware.
Combustors are usually operated at near lean blowout limits. Lean premixed combustion has a high potential of NO x
abatement. A stable combustion of lean mixtures with low flame speeds is therefore necessary in order to obtain
emission levels below 10 ppm [11]. Yet, as the adiabatic flame temperature is lowered, carbon monoxide emissions
tend to increase. An optimum operating point needs to be sought so as to guarantee both low nitric oxide and carbon
monoxide emissions.
Finally, the most important issues with SGH (syngas and hydrogen) fuel combustion are identified as; ability of the
combustor to burn variable mixtures of syngas without necessitating a design change, flame flashback, auto-ignition
phenomena, combustion dynamics such as thermo-acoustic instability, near-lean-limit flame instability related to issues,
feed system coupling, flow interaction and other unsteady fluid mechanics phenomena.
In order to investigate fuel flexibility issues and associated challenges experimentally a laboratory scale test rig is
built. The effect of hydrogen enrichment of methane fuel is investigated. Conventional and aerodynamic flame
stabilization techniques are also discussed.
Fig. 1 Borghi-Peters regime diagram for premixed
flames (note that conventional gas turbine
combustors operate within the thin reaction zones
regime).
2. Experimental setup
Experimental setup includes the combustor, gas bottles, metering systems, and the instrumentation. The combustor
system consists of the combustor shell, the inlet fuel and air-delivery system, and the premixing section defines the
dump plane. Quartz windows enable optical access into the main recirculation zone above the dump plane. At the exit
there is a cylindrical stainless steel shell followed with as conical constriction section where the exhaust flow diameter
drops down to 12 mm. Combustor is shown in Figure 2.
On the setup there are ports to enable pressure transducer mounting and there is also a port to enable gas sampling.
The basic design of the fuel-air premixing section represents a generic configuration with characteristic features similar
to industrial gas turbine systems where the fuel is injected into the swirling cross flow and mixes within a downstream
distance before reaching the dump plane. Combustor is operated up to a power rating of 20 kW.
©FORMATEX 2013
947
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
Hydrogen and methane gases are individually supplied from compressed tanks and mixed within a manifold prior to
combustor inlet. Their flow rates are controlled by separate mass flow meters. Mass flow rates are adjusted separately in
order to achieve the desired fuel composition. Air necessary for combustion is supplied from a compressor. Volumetric
airflow rate is measured by a rotameter, and a pressure gauge. Further details on the experimental setup can be found in
[12].
Fig. 2
Overall view of the combustor used for experimental study.
Combustion air is fed through a swirl vane (Figure 3). The blades on the swirl vane incorporate a swirling motion to
the fluid flow. Swirl provides stabilization at the dump plane and facilitates the entrainment of fuel jets within the cross
flow at the pre-mixer. The design is modular such that it can incorporate both a conventional swirl vane and a low swirl
vane.
a)
Fig. 3
b)
Conventional (a) and low swirl (b) vanes.
This flow then interacts with the by-pass stream at the center. By-pass jets are intentionally made with small diameter
such that the flame does not flashback. When the flow reaches the dump plane conservation of angular momentum
causes the flow to expand and separate. The periphery of the central separated flow is covered with a low intensity
swirling flow that gets weaker in the streamwise direction. A linear drop in the mean flow velocity occurs due to the
mean flow gradient. These set the conditions for the propagation of a premixed turbulent flame. Flame is stabilized at
the location where the local flow velocity is of the same magnitude but in opposite direction to the local flame
propagation speed [13-14]. Furthermore since the turbulence intensity is directly proportional to the flow velocity a
feedback loop is formed. Due to this mechanism acceleration and deceleration of the flame under different load
conditions is assured [15].
and
are the
The calculation of the swirl number can be performed with the following formula (Eq. 1).
corresponding radii of the center channel and the burner. The parameter denotes their ratio (Eq. 2). Similarly,
and
are the mass fluxes through the center channel and the swirl annulus. Therefore the parameter
is termed as the
by-pass ratio (Eq. 3). Note that when there is no-bypass flow through the center (i.e.
), then this situation
corresponds to a traditional high-swirl burner [13]. The introduction of the by-pass stream provides an additional
948
©FORMATEX 2013
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
degree-of-freedom for the designer. Hence not only the desired swirl number can be obtained by altering the flow split,
but also the desired divergent flow profile can be achieved [13].
(1)
(2)
(3)
3. Effects of hydrogen enrichment in conventional swirl stabilized configuration
Swirl stabilization offers unique characteristics due to the flame stability, high combustion performance and efficiency
both for premixed and non-premixed flames. Present gas turbine combustors and industrial systems alike both utilize
strong swirl [16]. In these systems the swirl component is strong enough to establish a recirculation zone within the
combustion chamber. Such a flow structure is shown in Figure 4. This re-circulation zone traps hot combustion products
within itself. Therefore, in traditional premixed combustion systems the re-circulation zone provides the heat source and
the free radical pool necessary for continuous ignition.
Fig. 4 Flow structure in a conventional swirl stabilized
) [12].
flame (ReD=19400, Sw=0.74,
3.1 Adiabatic Flame Temperature and Flame Speeds
Following the assumptions of Yu et al. [17] an equivalence ratio is defined as follows (Eq. 4). This equation implies that
the hydrogen in the blend is completely oxidized and the remaining oxygen is used to burn the methane content. This is
a reasonable assumption since the hydrogen oxidation proceeds much faster than methane oxidation.
φ=
CF C A − CH ( CH C A ) st 
( CF CA )st
(4)
Figure 5 shows the effect of hydrogen enrichment on adiabatic flame temperatures as a function of equivalence ratio
and hydrogen fraction. Adiabatic flame temperatures increase with increasing hydrogen percentage. Note that this
change is quite significant especially for lean mixtures. Figure 6 on the other hand illustrates the effect of hydrogen
enrichment on flame propagation speeds. Addition of hydrogen into the mixture significantly enhances the laminar
flame speeds. This in turn can have profound effects in terms of flame stabilization.
©FORMATEX 2013
949
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
3.2 Lean Blowout Measurements
Contemporary land based gas turbine engines are operated near their lean blow-off limits due to emissions
considerations. Here blow-off limits for the laboratory scale combustor are identified with respect to the particular
choice of fuel or fuel composition for that matter. Effects of two possible blowout mechanisms were discussed; one
approach based on front propagation, the other based on a well stirred reactor approximation. As fuel flexibility is an
issue and different fuel blends are used, it is possible that burning occur in different regimes.
2400
2200
Tad (K)
Fig. 5 Adiabatic flame temperatures of methane/hydrogen
mixtures as a function of equivalence ratio and hydrogen
fraction (hydrogen percentage by volume).
% 0 H2
%20 H2
%40 H2
%60 H2
%80 H2
2000
1800
1600
0.5
0.75
Φ
1
1.25
1.3
% 0 H2
% 20 H2
% 40 H2
% 60 H2
% 80 H2
Laminar Flame Speed (m/s)
1.2
1.1
1
Fig. 6 Laminar flame speeds of methane/hydrogen
mixtures (hydrogen percentage by volume).
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.6
0.8
1
Equivalence Ratio
1.2
1.4
Hydrogen enrichment considerably extends the lean blowout limits of the methane fuel as evidence to this argument
is shown in Figure 7. These effects regarding the extension of lean blowout limits through hydrogen addition are
, where α
consistent with the observations reported by [18-20]. Blowout data scales with an empirical parameter
is the hydrogen volume fraction. The higher the volumetric flow rate (hence velocity) the more the amount of hydrogen
added to methane in order to keep the reaction going. It is observed that for hydrogen enriched methane lean blowout
equivalence ratio is not fixed it both depends on the extent of enrichment and also upon the flow rate.
In flamelet combustion regime blowout occurs when the local flame speed is less then the oncoming fluid velocity
everywhere in the flame. Thus the stabilization mechanism for flamelet-like combustion is flame front propagation [20].
For flamelet combustion a loading parameter is defined as follows (Eq. 5).
950
©FORMATEX 2013
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
L=
ST
(5)
U
Turbulent flame speed is often expressed as in terms of laminar flame speed multiplied by a function which depends
both on turbulence intensity and geometry. For all conditions tested in the laboratory combustor the geometry is fixed
so those can be factored out when correlating the blow out behavior. Assuming turbulence intensities are similar as a
first order approximation the loading parameter can be expressed in terms of the laminar flame speed of the fuel mixture
[20]. Re-expressing the loading parameter in terms of laminar speed as in Eq. 6.
L=
SL
(6)
U
0.8
0.6
0.55
0.5
0.45
0.4
SL/U
0.35
0.3
0.25
0.2
0.15
0.1
0.7
Reactor Loading Parameter L2
0% H2
10% H2
20% H2
30% H2
40% H2
50% H2
0.6
% 0 H2
% 30 H2
% 40 H2
% 50 H2
0.5
0.4
0.3
0.2
0.1
0.05
0
0.3
0.35
a.
0.4
0.45
0.5
0.55
0.6
Equivalance Ratio at Blowout
0.65
0
0.3
0.7
0.4
0.5
0.6
Equivalance Ratio at Blowout
0.7
b.
Fig. 7 Relationship between blowout equivalence ratio and flamelet based loading parameter.
Figure 7a shows the relationship between lean blowout equivalence ratio and flamelet based loading parameter. Data
,
as well as a wide
for blowout is recorded in a wide range of operating conditions
range of fuel compositions (from pure methane to 50% methane/50 % hydrogen). As it can be seen from this figure data
points are scattered along a single line with a correlation coefficient of R2=0.69.
Chemical time scales for hydrogen and methane oxidation are quite different. Therefore considering the turbulence in
practical combustors combustion process can indeed cover a wide spectrum of Damköhler numbers. Another reactor
loading parameter L2 based on a well stirred reactor approach can be defined as well (Eq. 7). Following the work of
Hoffman et al. [21] azimuthal velocity component U and combustor diameter D are used as the appropriate scaling
parameters considering the fluid dynamic effects. If the well-stirred reactor based combustion is dominant then the
appropriate loading parameter L2 is defined as per the following equation. Here α denotes the thermal diffusivity of the
reactant mixture.
L2 =
αU
(7)
SL2 D
Results for reactor based correlation are shown in Figure 7b. For this parameter L2 correlation is weaker with respect
to the previous one. The correlation coefficient is only R2=0.41. Especially for low hydrogen content mixtures one
observes that there is a very large scatter in the data. This in turn points out that well-stirred reactor based loading does
not offer a good explanation for methane rich mixtures. On the other hand, one can see that for hydrogen rich mixtures
data seems to correlate well with this loading parameter. Therefore, no single assumption could entirely explain the lean
blowout behaviour of the combustor with methane hydrogen mixtures and tightly collapse all data points on a single
line. Flamelet based loading parameter on the other hand offers a much better correlation for low hydrogen content
mixtures. For hydrogen rich mixtures on the other hand there is a better correlation regarding the reactor based loading
parameter.
©FORMATEX 2013
951
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
3.3 Pressure, Heat Release and Flashback Measurements
A number of piezo-electric pressure transducers are mounted along the combustor wall measures dynamic pressure
variations in the combustor. So as to examine the waveform of the combustion instability pressure inside the reactor is
to be measured at several stations along the entire length of the combustor.
The CH/OH radical light intensity is recorded using a photodiode looking at the flame equipped with an appropriate
band-pass optical filter. Photodiode reading is taken as a measure of integral heat release fluctuations in the main
reaction zone of the reactor. Figure 8 shows the Fourier spectra of CH light, flashback and pressure signals at ReD=7200
and Φ=0.7. All three exhibit oscillatory behaviour as governed by the acoustic modes of the combustor. These modes
depend on the boundary conditions at the inlet and outlet. Interestingly a sudden shift in the dominant frequency occurs
after certain hydrogen content within the fuel is exceeded. In Figure 8, one can identify two distinct regions one with
high frequency and the other one with low frequency. With increasing hydrogen content a low frequency flashback
mode is observed. This behaviour occurs within a wide range of load conditions as evidenced by Figure 9.
Another way fuel composition affects the feedback route between heat release and pressure fluctuations within the
thermo-acoustic loop is the change that occurs when the flame center of heat release is altered due to change in flame
speed. As the flame speed increases with the addition of hydrogen the flame becomes shorter and the distance to the
flame center of heat release decreases. This yields to a reduced convective time for the equivalence ratio perturbations
which can cause a stable flame become unstable and vice versa [22]. Figure 10 shows the change in flame center of
mass with varying fuel composition while keeping cold flow velocity constant to factor out the aerodynamic effects.
These are time averaged two dimensional OH* chemiluminescence images captured with an intensified CCD camera
equipped with an appropriate band pass filter. OH* radical is a good indicator for heat release. One can notice the
change in the location between two flow conditions tested. On the right hand side with 50% methane most of the heat
release takes place in spatial locations closer to the dump plane.
Pressure
Heat Release
Flashback
60
80
40
60
No flashback
45
70
40
60
35
30
50
55
50
Pressure (dB)
Pressure (dB)
70
60
Pressure
Heat Release
Flashback
156 Hz
80
50
164 Hz
40 Hz
Flashback mode
90
Heat Release, Flashback (dB)
86 Hz
Heat Release, Flashback (dB)
90
30
50
25
40
0
100
200
300
Frequency (Hz)
400
20
500
40
0
100
200
300
Frequency (Hz)
400
a.
Fig. 8
20
500
b.
Pressure, heat release and flashback spectra, ReD=7200, Φ=0.7 (a. Pure Methane, b. %50 H2 by volume).
Another important phenomenon in the study of pre-mixed flames is flame flashback. Flashback can be triggered by
acoustic velocity fluctuations and is facilitated by higher flame speeds. Different fuels therefore have varying degrees of
susceptibility to flashback. This is not only due to a change of flame speed other factors are also present. For example a
change in the flame height alters the location of heat release or a change in the flame temperature affects the acoustics
by shifting the dominant mode.
Flame front movement and thermo-acoustic phenomena are in fact closely coupled to one another though the
dynamic equations. Velocity fluctuations cause the flame to move. This flame movement causes the spatial location of
heat release to move thus shifting the instantaneous phase between pressure and heat release.
A sequence phase locked of images is recorded showing the flashback cycle. This cyclic behaviour can be seen in the
CH chemiluminescence images of Figure 12. Initially flame cone makes a steeper angle with the dump plane in order to
balance the increase in the oncoming fluid velocity. Shortly afterwards flame departs from its attachment point and
continues to propagate upstream. At one point flame reaches its maximum propagation distance. As the total velocity
becomes positive it pushes the flame tip back to its attachment point. This behaviour is repetitive due to the nature of
velocity fluctuations occurring inside the combustion chamber.
Figure 11 depicts the cycle of events leading to periodic flame flashback. The attached wedge shaped flame moves
upstream due to a flow reversal shortly following pressure build-up. Vertical arrows pointing upwards and downwards
indicate total flow velocity and flame speed respectively. Whenever the flame is detached from the injector tip the flame
anchoring point (tip of the wedge shaped flame) moves with a velocity which is the sum of the flame speed and fluid
952
©FORMATEX 2013
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
velocity. As it is seen from the figure at the mid-cycle flame is entirely inside the pre-mixer. As the cycle progresses
and the total fluid velocity attains a positive value that beats the flame speed which always points towards the reactants
side flame front recovers its shape and re-attaches to the center body tip.
120
Fig. 9 Dominant frequency (Hz) with respect to
fuel composition and equivalence ratio
(ReD=7200).
Frequency (H z)
100
80
60
40
20
1
0.9
0.8
Eq. Ratio
0.7
0.6
0
0
10
20
% Hydrogen
30
40
50
a. Pure Methane
b. 50% H2 by volume
Fig. 10 OH* chemiluminescence images demonstrating effect of fuel composition on the distance to the flame center of mass at a
fixed flow rate and equivalence ratio (ReD =7200 , Φ=0.7, Sw=0.74).
3.4 Emissions
Figure 13 demonstrates the effect of hydrogen enrichment on nitric oxide emission index. Emission index is defined as
the grams of pollutant generated per kg of fuel burnt. This non-dimensionalization offers an unbiased comparison. It is
observed that EINO values rises monotonically with increasing hydrogen fraction. This effect is due to increased flame
temperatures.
Figure 14 shows the correlation between adiabatic flame temperature and nitric oxide emissions index (EINO) in a
wide parameter space including fuel variability effects. Note that for the flow condition tested residence time changes
by almost a factor of three. Residence time is one of the determining parameters for the equilibrium concentration of
species. However as the points scatter along a line depending on the adiabatic flame temperature one can argue that
residence times are longer than formation timescales and the associated effects are relatively smaller. Points on the
scatter plot are color coded according to the binary fuel mixture composition. Holding all the other parameters constant,
and increasing hydrogen volume fraction within the fuel blend, yields an increase in the nitric oxide emissions index. A
possible explanation for such an affect is the increased abundance of hydroxyl radicals within the radical pool. Higher
hydroxyl concentrations can increase the forward reaction rates in the nitric oxide formation mechanism where
©FORMATEX 2013
953
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
hydroxyl takes part in. Figure 14 suggests that the main scaling parameter is the adiabatic flame temperature for oxygen
based EINO. It can be concluded that the extended Zeldovich mechanism (thermal NOx) is the dominating pathway to
NO formation.
Fig. 11 An illustration showing the cycle of
events regarding flashback.
0 degrees
45 degrees
90 degrees
135 degrees
Fig. 12 Phase locked CH radical images
demonstrating a flashback cycle (ReD= 6600,
φ=0.7, 40 % hydrogen by volume).
100
80
60
40
20
0
180 degrees
225 degrees
270 degrees
315 degrees
4. Aerodynamic flame stabilization
Low swirl aerodynamic flame stabilization, as introduced by Cheng [13], is a rather new concept. Its working principle
is based on the propagation of the premixed flame front. Premixed flames create a continuous reaction due to their selfpropagation ability. The flame speeds depends on the concentration of the reactants, thermodynamic conditions and
turbulence intensity. In low swirl combustion in order to hold turbulent fast propagating flames a diverging flow field is
utilized [14]. A diverging flow field is established under conditions such as vortex breakdown when the swirl strength is
low. Therefore in a low swirl combustor the flame is stabilized aerodynamically. This is the fundamental difference
between low swirl combustion and traditional techniques.
The following discussion provides an insight to this problem and enables one to reach certain conclusions. The axial
velocity at the flame anchoring location is provided by the following equation (Eq. 20) [13].
(8)
954
©FORMATEX 2013
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
In the literature it has been demonstrated that for low swirl flames the turbulent flame speed demonstrates a linear
correlation for a wide range of fuels including methane, propane, ethylene and diluted hydrocarbons [23-24]. This
relationship is provided below through Eq. 9.
0.7
φ=0.7
EINO (gNO/kgfuel)
0.6
0.5
0.4
0.3
0.2
0.1
10
20
30
40
50
% Hydrogen Volume Fraction
Fig. 13 Effect of hydrogen enrichment on emissions index (ReD=4870, Sw=0.74).
(9)
and substitutes the above correlation (Eq. 9) for
Hence if one divides this expression (Eq. 8) by the axial velocity
the turbulent flame speed
, then the following expression would eventually emerge (Eq. 10) [15]. Now this
expression (Eq. 10) can be analyzed term by term. Note that at the left hand side
is invariant due to flow
similarity, furthermore
is asymptotic at large , which is often the case in gas turbine combustors. At the
right hand side the term
is rather small; the other term also remains nearly constant. As a consequence, flow
field similarity and turbulent flame speed correlations do explain why the flame can be stabilized aerodynamically and
remains stationary through a wide range of velocities and equivalence ratios [13-15].
(10)
Fig. 14 Relationship between adiabatic flame
temperature and emissions index in a wide
range of load conditions (0.4<φ<1.4,
2900<ReD<8200, Sw=0.74).
0.25
EINO (O2 Based)
0.2
%0 H2
%10 H2
%20 H2
%30 H2
%40 H2
%50 H2
0.15
0.1
0.05
0
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
T/Tref
1
1.05 1.1 1.15 1.2
©FORMATEX 2013
955
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
5. Discussion and conclusion
Hydrogen-enriched confined methane combustion is studied in a laboratory scale premixed combustor. Whenever the
equivalance of hydrocarbon fuels is of concern, the gas turbine industry utilizes a parameter called Wobbe index
defined as per Eq. 11. Here HHV indicates the higher heating value. For pure methane this index is 12.7 whereas it 11.5
for pure hydrogen.
(11)
Hydrogen enrichment both increases the adiabatic flame temperatures and the flame speeds. Since the flame
propagation speed of hydrogen is much faster that of methane shifts in the combustion regime were observed. Hydrogen
enrichment increases nitric oxide emissions due to rising adiabatic flame temperatures, however at the same time this
also enables very lean mixtures to be burnt (thus reducing flame temperatures) and the overall effect becomes a
reduction in the nitric oxide emissions. It is thought that the thermal path (Zeldovich mechanism) is dominant in the
production of nitric oxides.
Correlating parameters for lean blowout (LBO), pressure amplitudes and emissions are examined. Two loading
parameters are examined to correlate the lean blowout results: one based on a flamelet approach the other based on a
well-stirred reactor approach.
Dominant acoustic mode can go sudden changes with increasing hydrogen content. Such behavior is observed for the
present combustor beyond 40% H2 by volume and suddenly a low frequency mode becomes dominant. Furthermore
with the conventional swirl vane flame flashback becomes a troubling issue with increasing hydrogen content within the
fuel mixture.
These conclusions altogether suggest that Wobbe index is not an appropriate scaling parameter when
methane/hydrogen mixtures are of concern. Due to profound changes in flame behavior, a novel design methodology is
needed in order to burn hydrogen rich fuels in premixed gas turbine combustors.
The divergent flow structure offered by the low swirl vane suggests that it would be the appropriate solution for
premixed methane/hydrogen fuel mixtures. The flame in this configuration would be stabilized purely by aerodynamics
and since at each load condition the aerodynamics would be such that an acceleration or deceleration would be balanced
by the oncoming flow structure the flame would burn without actually touching anywhere inside the combustor (i.e. a
center body would not be necessary for flame holding). Thus the disparity between flame speeds of methane and
hydrogen would not be of concern. Moreover, this approach would offer a simple retrofit for the existing gas turbines.
Replacing the conventional swirler with a low swirl vane shall fix the problem and offer a wide range of fuel variability.
This aerodynamically stabilized hydrogen enrichment approach then indeed can pave the road towards a full hydrogen
economy in the future.
Fig. 15 Diverging flow pattern past a low swirl vane,
ReD=50700, Sw=0.62 (velocities non-dimensionalized
with respect to maximum axial velocity).
Acknowledgements The support by Turkish Scientific and Technical Research Council (TÜBİTAK) through grant number
109M426 is gratefully acknowledged.
956
©FORMATEX 2013
Materials
and processes for energy: communicating current research and technological developments (A. Méndez-Vilas, Ed.)
____________________________________________________________________________________________________
References
[1] Smoot LD, Smith PJ. Coal Combustion and Gasification. Plenum Press; 1985.
[2] Tuncer O. Active Control of Spray Combustion. PhD Dissertation, Louisiana State University, Baton Rouge, LA, 2006.
[3] Calvetti S, Carrai L, Cecchini D. Esecuzione di Prove di Co-Combustione di Gas Naturale e Syngas da Biomassa su un
Combustore DLN per Turbogas, Technical report, Enel Produzione, ENELP/RIC/RT/-2001/258/0-IT+RT.RIC.PI, Pisa, Italy,
2001.
[4] Bonzani F, Pollarolo G, Ferrante A. Ansaldo V94.2K Gas Turbine Burner Performances Operating with Steelworks Process GasNatural Gas Fuel. Power-Gen Europe, 2000.
[5] Chen RH. Some Characteristics of NOx Emission of Turbulent Nonpremixed Hydrogen Air Flames Stabilized by Swirl Generated
Flow Recirculation. Combustion Science and Technology, 1995;110:443-460.
[6] Chen JY. Development of Reduced Mechanisms for Numerical Modeling of Turbulent Combustion. Workshop on Numerical
Aspects of Reduction in Chemical Kinetics, CERMICS-ENPC Cite Descartes-Champus sur Marne, France, 1997.
[7] Borghi R. On the structure and morphology of premixed flames. Recent Advances in Aerospace Sciences, Plenum Press, New
York. 1985;117-138.
[8] Peters N. Laminar flamelet concepts in turbulent combustion. Proceedings of 21st Symposium (International) on Combustion. The
Combustion Institute, Pittsburgh, 1986:1231-1256.
[9] Mariotti M, Tanzini G, Faleni M, Castellano L. Sperimentazione di Fiamme di Idrogeno a Pressione Atmosferica in un
Combustore per Turbogas con Iniezione di Inerti. Technical report, Enel Produzione, Pisa, Italy, ENELP/RIC/RT/-2002/0063,
2002.
[10] Tomzcak H, Benelli G, Carrai L, Cecchini D. Investigation of a Gas turbine Combustion system Fired with Mixtures of Natural
Gas and Hydrogen, IFRF Combustion Journal, Article Number: 200207, 2002.
[11] Kiesewetter F, Hirsch C, Fritz M, Kroner M, Sattelmayer T. Two-Dimensional Flashback Simulation in Strongly Swirling
Flows. ASME Paper No: GT2003-38395, 2003.
[12] Tuncer O, Kaynaroglu B, Karakaya MC, Kahraman S, Çetiner-Yıldırım, Baytaş C. Preliminary Investigation of a Swirl
Stabilized Premixed Combustor. Fuel, doi: 10.1016/j.fuel.2012.11.085.
[13] Cheng, R. K., Littlejohn, D. Effects of Combustor Geometry on the Flowfields and Flame Properties of a Low-Swirl Injector.
ASME Turbo Expo 2008, Berlin, Germany, ASME Paper No: GT2008-50504, 2008.
[14] Cheng RK, Littlejohn D, Nazeer WA, Smith KO. Laboratory Studies of the Flow Field Characteristics of Low-Swirl Injectors
for Application to Fuel-Flexible Turbines. Journal of Engineering for Gas Turbines and Power, 2008;130:21501-21511.
[15] Littlejohn D, Cheng RK, Noble DR, Lieuwen T. Laboratory Investivations of Low-Swirl Injector Operating with Syngases.
ASME Turbo Expo 2008, Germany, ASME Paper No: GT2008-51298, 2008.
[16] Syred N, Beer JM. Combustion in Swirling Flow: A Review. Combustion and Flame, 1974;23:143-201.
[17] Yu G, Law CK, Wu CK. Laminar Flame Speeds of Hydrocarbon Plus Air Mixtures with Hydrogen Addition, Combustion and
Flame, 1986;63:339-347.
[18] Schefer RW. Reduced Turbine Emissions Using Hydrogen Enriched Fuels. Proceedings of the 2002 U.S. DOE Hydrogen
Program Review, NREL/CP-610-32405, 2002.
[19] Guo H, Smallwood GH, Liu F, Ju Y, Gulder O. The Effect of Hydrogen Addition on Flammability Limit and NOx Emission in
Ultra Lean Counterflow CH4/Air Premixed Flames. Proceedings of the Combustion Institute, 2005;30:303-311.
[20] Zhang Q, Noble DR, Meyers A, Xu K, Lieuwen, T. Characterization of Fuel Composition Effects in H2/CO/CH4 Mixtures Upon
Lean Blowout. ASME Paper No: GT2005-68907, 2005.
[21] Hoffman S, Habisreuther P, Lenze B. Development and Assesment of Correlations for Predicting Stability Limits of Swirling
Flames. Chemical Engineering and Processing, 1994;33:393-400.
[22] Hemchandra S, Shreekrishna, Lieuwen T. Premixed flame response to equivalence ratio perturbations. 43rdAIAA/ASME/ASEE
Joint Propulsion conference, Cincinnati, Ohio, AIAA Paper No: AIAA-2007-5656.
[23] Shepherd IG, Cheng RK. The Burning Rate of Premixed Flames in Moderate and Intense Turbulence. Combustion and Flame,
2001;127:2066-2075.
[24] Shepherd IG, Kostiuk LW. The Burning Rate of Premixed Turbulent Flames in Divergent Flows. Combustion and Flame,
1994;96:371-380.
©FORMATEX 2013
957