SPECIAL REPORT: HYDROGEN
HYDROGEN
COMBUSTION
Solving The Challenge Of Lean Premix Combustion
With Highly Reactive Fuels
BY JEFFREY GOLDMEER
T
he power generation industry is in transi- accessory systems that can accommodate the
tion with increasing momentum to reduce increased volumetric flow. Increased reactivity
carbon emissions. Each year more renew- is another challenge. Figure 1 shows relative
able-based power is generated, changing the reactivity as a function of inverse blow-off time,
traditional electrical grid model. However, the which is taken to be a characteristic chemical
grid still requires firming with dispatchable
assets. In the future, these assets may need to
have reduced or zero carbon emissions.
Transforming gas turbines into low or
zero-carbon emitting systems can be done
through pre- or post-combustion options.
Pre-combustion choices include the use of
hydrogen, renewable or synthetic methane, or
biofuels. Post-combustions options include car-
bon capture and oxy-fuels. As hydrogen is
both a fuel and an energy carrier, it is being
viewed as a potential critical element for
decarbonization of many sectors, including
power generation.
Fuels containing hydrogen have been used
extensively in the power generation industry for
decades. However, there are multiple chal-
lenges due to differences in fundamental char-
acteristics (Table 1). Hydrogen has a higher
flame temperature than methane, which may
lead to increased NOx production. Hydrogen’s
lower heating value (on a volumetric basis) is
1/3 the value of natural gas, requiring fuel Figure 1: Fuel reactivity. Courtesy of GE Gas Power.
CHARACTERISTIC UNITS METHANE HYDROGEN
Formula CH4 H2
Molecular weight 16 2
Lower / Upper flammability limits % 4.4 / 17 4 / 75
Adiabatic flame temperature °F (°C) ~3,565 ~4,000
Lower heating value MJ/Nm3 (BTU/scf) 35.8 (911.6) 10.8 (274.7)
Lower heating value MJ/kg (BTU/lb) 50 (21,515) 120 (51.593)
Table 1: Characteristics of methane and hydrogen
14 www.turbomachinerymag.com November/December 2020 • Turbomachinery International
SPECIAL REPORT: HYDROGEN
reaction time. This is plotted against the mod-
ified wobbe index. From this graph, it is seen
that high hydrogen fuels have a relative reac-
tivity that is ~100x faster than natural gas fuels.
Greater reactivity leads to increased flame
speed (defined as the velocity a flame will prop-
agate upstream into unburned fuel). Hydrogen’s
f lame speed is an order of magnitude faster
than methane (Figure 2). Even blends of
hydrogen and natural gas will exhibit increased
flame speed; a 50% blend of hydrogen and nat-
ural gas has a flame speed at least twice that of
100% methane. Higher flame speed increases
the r isk that the f lame could propagate
upstream into the premixer. If the flame is able Figure 2: Flame speeds of hydrogen blends
to anchor and stabilize inside the premixer, this
is known as flame holding. Both situations can
lead to combustion hardware distress and even
fuel nozzle damage (Figure 3).
To mitigate some of the challenges of operat-
ing with hydrogen and similar low heating value
fuels (i.e. syngas, steel mill gases, refinery waste
gases, etc.), gas turbines were configured with
diffusion flame combustion systems. These com-
bustors had physically different geometries that
mitigated the risk of flashback and flame hold-
ing. However, these systems operate at combus-
tion conditions that allow near peak-to-peak Figure 3: An example of damage caused by a flame holding event on a dry low NOx
NOx formation (Chemically, these are stoichio- (DLN) fuel nozzle. Damage can be seen along the central hub and on the tip of the
nozzle, Source: GE Gas Power
metric reactions where fuel and air are present
in balanced chemical proportions; these reac-
tions also have an equivalence ratio of one) (Fig-
ure 4). Depending on the fuel and the combus-
tion system, this could result in NOx emissions
in the range of 200-600 ppm. A typical mitiga-
tion for high NOx emissions is injection of a dilu-
ent (water, steam, nitrogen) into the combustor,
but this may result in reduced performance.
A path to resolving this challenge is shift-
ing to equivalence ratios less than one; this is
where dry low emissions (DLE) and dry low
NOx (DLN ) combustion systems operate.
However, traditional DLE and DLN systems
are not used with hydrogen due to concerns of
flashback and flame holding.
One solution was to develop a different type
of premixed combustion system with modified Figure 4: Premixed flame compared to lean premixed flame. Source: GE Gas Power
geometries for swirl-based DLE and DLN sys-
tems. An alternative approach to hydrogen
combustion that mitigates the risk of flashback
is lean direct injection (LDI). As the name
implies, LDI-based combustors operate in a fuel
lean mode and achieve rapid mixing of fuel and
air using hundreds or thousands of small-scale
mixers. The scale of these injectors is such that
the bulk velocity of the gas exiting the injector
is higher than the f lame speed of the fuel,
reducing the potential of f lashback or f lame Figure 5: Multi-tube mixer concepts, left to right from NASA, Hitachi, Parker Hannifin,
holding. In addition, the diameter of these holes and GE Gas Power
Turbomachinery International • November/December 2020 www.turbomachinerymag.com 15
SPECIAL REPORT: HYDROGEN
may be smaller than the quench distance of the
f lame, mitigating the potential risk of f lame
holding (The quench distance for a flame is the
distance between two plates, or the diameter of
a tube, that will quench a f lame when it
attempts to propagate through the geometry.
Physically, the walls of the channel or tube wall
conduct enough heat away from the flame that
it is extinguished locally).
Numerous studies examine LDI variations
and configurations, including co-flow mixing,
swirl-based mixing, and jet in cross-flow mix-
ing (Figure 5). All feature a number of small
diameter fuel injectors. Due to their scale, they
require more intricate methods of delivering
fuel and air. The NASA mixer uses a system
in which air flows axially into the multi-tube Figure 6: NASA low emissions LDI hydrogen combustor assembly
assembly and through the 25 injection ele-
ments (Figure 6). The hydrogen flows into a
fuel plenum in the assembly, and is injected
axially into two locations in each element, cre-
ating a jet in cross flow mixing system.
The jet in cross-flow ( JICF) is classic fluid
mechanics mixing process where a jet is injected
nominally perpendicular to a primary flow. In
Figure 7, the jet breaks up and starts to mix as
it is sheared by the primary flow.
Combining these concepts, a hybrid LDI/
JICF combustion system utilizes a very large
number of injectors as well as a jet in cross-flow
Figure 7: Visualizations of jet in cross-flow mixing; (a) smoke injection in a cross flow;
to create a premixing region inside each injec-
(b) numerical simulation of injection jet in cross-flow mixing
tor (Figure 8). In this mixer, air flows through
numerous straight tubes in a parallel array with
a fuel plenum that surrounds the tubes. Like the
NASA mixer, fuel flows into the mixer through
a ring manifold on the outer diameter and is
injected radially into the air f lowing axially
through small-diameter holes in the tube walls.
The distance from the injection point to the
tube exit plane may vary depending on the fuel
and engine conditions. As these tubes have
small diameters, bulk air velocity may be
greater than the flame speed of the fuel to avoid
flashback, including velocities greater than the
flame speed for fuels with significant hydrogen Figure 8: Jet in cross-flow multi-tube mixers. Courtesy of GE Gas Power.
content. Even though generating these high
velocities can impact performance, due to the
short axial length of these mixers, they are able
to achieve a relatively low pressure drop.
An advantage of the multi-tube mixer that
features many small elements is the scalability
of the design to any size nozzle or combustor
with limited impact on performance. This has
allowed initial concept systems to be scaled up
to full-scale gas turbine hardware. This has
been demonstrated by multiple research orga-
nizations and gas turbine manufacturers who Figure 9: Full-scale multi-tube combustion systems used for lab testing and/or
developed full-scale prototypes for testing and commercial operation (a) Mitsubishi Power, (b) Kawasaki Heavy Industries,
commercial operation (Figure 9). (c) GE Gas Power
16 www.turbomachinerymag.com November/December 2020 • Turbomachinery International
SPECIAL REPORT: HYDROGEN
GE’s new multi-tube combustion system is
known as the DLN 2.6e (Figure 10). Although
optimized for operation on natural gas, it still
utilizes the LDI/JICF physical geometry and
mixing concepts that provided improved oper-
ability on hydrogen. Combustion tests demon-
strated that this system can operate on blended
hydrogen and natural gas with up to 50% (by
volume) hydrogen.
D r. J e f f r e y G o l d m e e r i s
Emergent Technology Director
for Decarbonization at GE Gas Figure 10: The GE DLN 2.6e combustion system plus an image taken from a combustion
Power, a provider of natural gas test operating on 100% natural gas. It is part of the standard configuration on GE’s
power technology, services, and 9HA and 7HA.03 gas turbines. The first 9HA with this combustion system is in
solutions. For more information, commercial validation at a site in Asia. The first 7HA.03 gas turbine will power Florida
visit GE.com/power/gas Power & Light Company’s (FPL) Dania Beach Clean Energy Center.
Turbomachinery International • November/December 2020 www.turbomachinerymag.com 17
