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

net/publication/267620101

The GT24/26 Combustion Technology and High Hydrocarbon (“C2+”) Fuels

Conference Paper · January 2005

DOI: 10.1115/GT2005-68799

CITATIONS READS

10 1,249

4 authors, including:

Oliver Riccius Felix Güthe

Alstom Technische Universität Berlin
14 PUBLICATIONS   292 CITATIONS    74 PUBLICATIONS   1,203 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Unimolecular decay of energy selected molecular ions View project

Fuel flexibility of gas turbine power plants View project

All content following this page was uploaded by Felix Güthe on 11 January 2015.

The user has requested enhancement of the downloaded file.

 Proceedings of ASME Turbo Expo 2005
Power for Land, Sea and Air
June 6-9, 2005, Reno, USA

GT2005-68799

THE GT24/26 COMBUSTION TECHNOLOGY

AND HIGH HYDROCARBON (“C2+”) FUELS

Oliver Riccius Richard Smith Felix Güthe Peter Flohr1

ALSTOM (Switzerland) Ltd

Segelhof
CH-5405 Baden-Dättwil (Switzerland)

ABSTRACT 1. INTRODUCTION

In 1998 ALSTOM launched a development program to
burn fuels containing high hydrocarbons in the GT24/26 gas
turbines. These fuels are found in different parts of the world as
for example in Brazil, Algeria or in the gulf of Bengal and
Thailand. After validation in the GT26 gas turbine in the
ALSTOM power plant in Birr (Switzerland) the combustion of
fuel gases with up to 16 Vol% of high hydrocarbons was
released in 2001. First field engines started to operate in 2001
and today 18 units are running commercially successful
burning high hydrocarbon fuel gases. The engines operate
without any hardware modifications compared against standard
natural gas fuel, however with an advanced engine control
system based on a fast gas composition monitoring system.
This achievement is particularly remarkable in view of the
sequential combustion system with a high pressure (30bar)
first combustor stage, and a high inlet temperature (1000°C)
second combustor stage.
The paper discusses fundamental aspects of burning high
hydrocarbon fuels and in particular its impact on the ignition
process. It outlines the careful development and validation
procedure which has been applied at ALSTOM. Finally,
evidence is given of today’s reliable engine operation with
200’000 operating hours on high hydrocarbon fuel supplies at
remarkably low emission levels.
Market requirement for fuel flexibility
The range of natural gases for gas turbine combustion has
become wider in recent years in terms of species composition,
Wobbe index (WI) and heating value (LHV). The gas
composition may include large fractions of rich gases or higher
hydrocarbons, which vary. Those higher hydrocarbons are
summarized as “C2+”, which is the sum of all mole fractions of
hydrocarbons with more than one C-atom. At the same time
todays gas supplies may carry a large and varying content of
inert gases, mainly CO2 and some N2. The variation of both,
C2+ and inerts is wider than anticipated and the changes occur
faster than previously known, and it is likely that this trend will
continue to rise in the future, see figure 1. For a customer in
Bowin/Thailand, table 1 gives a more detailed analysis of
typical fuel compositions. It contains inert gases of up to 19%
and higher hydrocarbons up to 16% while the Methane content
is as low as 65%. The rich fuels as Ethane, Propane and Butane
are combusted in fractions of 7 to 9%, 1.4 to 4.5% and 0.5 to
2% respectively.
The gas composition for a power plant may in fact change
rapidly in time which asks for robust solutions where gas-
dependent hardware modifications may severely limit the
operational flexibility for an engine. In figure 2, a time-record
of a gas chromatograph illustrates this issue where variations of
C2+ between 7 and 14Vol% have been observed.

1
peter.flohr@power.alstom.com

1 Copyright © #### by ASME

 GT26
GT26 Field
Field Engines
Engines with
with Standard
Standard Fuel
Fuel Gas
Gas The reheat combustion system
52
ALSTOM’s large power gas turbines, the GT24/26 family, use
48 a sequential combustion system (figure 3). The compressed air
44 is heated up in a first annular combustion system, after which it
expands in a 1-stage high pressure turbine. The exhaust air,
LHV [MJ/kg]

GT26
GT26 Field
Field Engines
Engines with
with Non-std
Non-std Fuel
Fuel Gas
Gas
40
containing residual oxygen from the lean premix combustion, is
36 then heated up again to maximal turbine inlet temperature after
32
which it expands in the low pressure turbine stages. This
unique gas turbine cycle allows to achieve higher part load
28 efficiency, as well as very low emission levels because
24
ALSTOM
ALSTOM GT26
GT26 Test
Test Center
Center Birr
Birr emissions are decoupled from the firing temperature in the first
0% 2% 4% 6% 8% 10% 12% 14% 16% 18% combustor stage.
C2+ [%]

Figure 1: Lower Heating Value versus higher hydrocarbon

SEV-Burner
content with a snap-shot of today’s market. The grey symbols
indicate typical gas compositions which are seen today in the
operating fleet, the red symbols indicate an additional focus for
fuel gases with high inert and low C2+ content.

Gas 1 Gas 2 Gas 3 Gas 4 Gas 5

CH4 65.32% 68.21% 71.50% 74.66% 70.02%
C2H6 8.68% 8.19% 7.33% 6.90% 8.94%
C3H8 4.47% 2.95% 2.35% 1.42% 2.98%
iC4H10 0.97% 0.61% 0.48% 0.25% 0.56%
Figure 3: cross-sectional view of the GT24/26 sequential
nC4H10 0.99% 0.63% 0.51% 0.27% 0.56% combustion system with two annular combustion chambers,
iC5H12 0.15% 0.11% 0.09% 0.05% 0.09% separated by a single-stage high pressure turbine.
nC5H12 0.19% 0.14% 0.12% 0.06% 0.12%
C6+ 0.20% 0.14% 0.13% 0.05% 0.11%
CO2 17.52% 17.42% 15.81% 14.62% 15.01% For variations in gas fuel composition, the development
N2 1.50% 1.63% 1.69% 1.74% 1.63% engineer faces a number of challenges, related to
sum 100.00% 100.00% 100.00% 100.00% 100.00% 1. mixing quality and emission levels
C2+ 15.66% 12.75% 10.99% 8.98% 13.35% High inert contents require higher fuel flow for a given power
inerts 19.02% 19.05% 17.50% 16.36% 16.64% output. Fuel distribution systems must be able to cope with
LHV [MJ/kg] 32.7 32.3 33.2 33.8 34.1 increased mass flows and pressures. Changes of the fuel
LHV [MJ/m3] 32.9 31.3 31 30.3 32.1
density [kg/m3] 1.01 0.97 0.93 0.9 0.94
injection must be considered as well. The momentum of fuel
WI [MJ/m3] 37.2 36 36.4 36.3 37.5 jets rises which impacts the mixing field. That in turn may
affect flame position and stability as well as the low NOx
Table 1: 5 typical gas compositions in pipeline gas in Thailand. characteristics of the engines.
2. flammability limits and flame flashback
The higher hydrocarbons typically increase the reactivity of the
fuel-air mixture which lead to increased flammability limits
(often advantageous) and higher flame propagation speed. The
second issue may result in a change in flame locality (usually
observed in terms of increased emission values), but in extreme
cases lead to flame flashback and combustor hardware damage.
3. auto-ignition
Spontaneous or auto-ignition of the fuel-air mixture is
enhanced by addition of C2+ species. While this issue is
generally of concern to all gas turbine manufacturers (and for
example reflected in a growing number of ASME papers in this
field), it becomes crucial to the design of a robust reheat
combustor because flame stabilization in the sequential
Figure 2: Changes of C2+ content over time in a typical gas combustor fully relies on spontaneous ignition.
turbine supply line measured with gas chromatographs.

2 Copyright © #### by ASME

Outline of paper
This paper highlights the fuel flexibility development project
for the GT24/26 gas turbines with the scope of theory, rig tests,
engine test, fleet implementation, and fleet experience.
The theoretical part should firstly characterize the customer
fuel gases and define model gases for rig and engine tests.
These model gases were required because it was not possible to
transport the customer fuels to our test rigs in sufficient
quantities. At the same time these mixtures must be designed
from components, which are readily available in Europe. This
work is summarized in section 2.
Component tests encompass combustion tests at atmospheric
and high pressure conditions as well as mixing studies in water
channels. The goal is to provide a first feedback regarding
combustor characteristics and operation concept at the single
burner level. Flame position and shape is monitored and
combustor hardware temperatures are being recorded while the
gas composition is varied. Finally, the engine tests provide the
completest validation about the engine combustion
characteristics with the new fuel gases, ahead of fleet
implementation. Combustion stability, pulsations, emissions,
performance as well as flame position and shape are to be
recorded while the gas composition must be variable. In order
to achieve that we have employed the GT26B test engine in
Birr, and installed a gas mixing plant to supply high C2+ and
high inert gases. Special equipment was to be prepared to
supervise the gas composition and monitor flame position and
metal temperatures at the combustor hardware. Based on the
test results the final operation concept had to be defined.
Combustor back-up hardware was to be evaluated as well. In
section 3 we summarize the effort and experience of the rig and
engine tests and also give a specific example of transferring rig
results to the engine.
The fleet implementation, achieved in 2001 and described in
section 4, must ensure safe and swift transfer of the test results
into the customer engines. In fact that scope could be reduced
to a revised operation concept and fast gas sensors. The starting
phase contained an extended commissioning, which includes
mappings through the operating window and monitoring of
temperatures and flame position. Combustor back-up hardware
has been verified as well. The implementation is accompanied
with a number of additional inspections during the first release
phase. Those inspections have confirmed the revised operating
concept and have now been reduced to the standard inspection
scope and intervals.
2. AUTO-IGNITION ANALYSIS
Background
The presence of minor hydrocarbon species has a profound
influence on the auto ignition behavior of fuel-air mixture.
Hydrocarbon species typically found in natural gas along with
methane include ethane, propane, butane and pentane and very
few higher alkanes as well as small contributions of partially
unsaturated alkenes. Especially the next higher members of the
homologous series of the alkanes ethane and propane are
important since they are relatively abundant. The sum of all
higher alkanes is usually referred to as C2+, without
differentiating between the different species. Their effect on
auto ignition time is however not equal. This has to be
understood in the evaluation of the burning properties of
natural gases as well as model fuels. In this paper we study
exemplarily the relative difference between ethane and propane
additions to a methane fuel.
Theoretical analysis of auto-ignition can be performed by the
SENKIN module of the program package like CHEMKIN,
based on a detailed reaction mechanism. We compare here a
number of mechanisms from published work and indicate
relative changes between them. As standard reference we use
the mechanism of the gas research institute (GRI3.0), which
seems to be a good compromise for all applications and is
usually not very far from experimental results. However higher
alkanes are apparently better modelled by Wang et al., due to
their more detailed C3 chemistry. Additionally, for low
temperatures it is important to account for peroxides to capture
low temperature chemistry. To estimate the effect of ethane and
propane under SEV conditions it has been shown that
combinations of these mechanisms seem to yield more reliable
results. A and B are two different kinetic schemes derived from
connecting the C3 chemistry and the low temperature C1
chemistry of several literature mechanisms.
In figure 4 the predicted ignition times for high temperature
conditions as in the SEV combustor are presented. The gas
mixture is standard Swiss gas; the difference between the
reaction mechanisms is moderate, however with a noticeable
trend for too short ignition times at high preheat temperatures
for GRI3.0. The difference of the mechanisms becomes more
significant for changes in gas composition as discussed below.
1.E-01
GRI3.0
A
1.E-02
τignition /s
B
1.E-03
1.E-04
1000 1100 1200 1300 1400
T(air) / K
Figure 4: predicted ignition times at reheat conditions for Swiss
gas with low C2+ content. Comparison of 3 different reaction
mechanisms. A and B are two different kinetic schemes derived
from connecting the C3 chemistry and the low temperature C1
chemistry of several literature mechanisms.
3 Copyright © #### by ASME
A reactor model of the reheat combustor
A chemical reactor network for the GT24/26 has been
developed which is fed directly from engine performance data
sheets. Pre-studies have shown that the ignition times depend
not only on temperature and pressure but also on the fuel type,
the inlet conditions and the composition of the exhaust gas (~
combustion air). That includes the concentrations O2 and H2O
as well as a number of other intermediates and radicals (like
OH, CH2O, etc..).
Using this approach it is possible to obtain the composition of
calculations even if calibrated against a reference operating
point.
1. To study the effect of increasing C2+ for a given ethane to
propane ratio (1:1) five different fuel mixtures were tested,
based on a reference natural gas with C2+ = 3% (Swiss gas) and
then increasing values of ethane / propane additive such that
heating value was kept constant.
1.2

the SEV mixture as precisely as possible. A change in fuel GRI

composition or cooling air is automatically accounted for. 1
6% C2+
In this network, the EV-flame is modeled by a PREMIX A
module followed by a SENKIN reactor (PFR) as post flame 0.8
zone. The number of parallel reactors can be chosen as a

τIgnition_C2+/τSwiss
B
compromise between computational time and precision. For
emissions prediction the degree of mixing is very important and 0.6
25% in τIgnition

needs to be calibrated from engine data. However if the

ignition time in the SEV is in the focus a smaller number of 0.4

reactors can be chosen.

After the EV-flame region the gas is mixed with more cooling
0.2
air and enters the high pressure turbine (HPT) which reduces
the temperature according to the pressure ratio and turbine
efficiency. After the HPT cooling is added and the gas enters a 0
0 5 10 15 20 25 30 35 40
PFR for 2 ms (approximate time between HPT and SEV fuel C2+ content

gas nozzle) to relax its radical concentrations towards their

equilibrium at the temperature and pressure after the HPT. Then
the gas stream (of EV exhaust and cooling air) is mixed with
fuel to enter the SEV combustor, which is simulated with
another PFR. For a given mass flow of fuel the flame
temperatures are calculated. But the program can also be used
to calculate the needed fuel mass flow for a given flame
temperature or turbine inlet temperature. This option is used to
model the engine operation concept when changing fuels to
compensate for differences in heating value. The output of the
module is a gas composition and an auto ignition time. Other
derived quantities, such as effective flame position and
thickness, are computed from the tool, as well, to help the
designer in optimizing the combustor flow distribution.
Effect of higher hydrocarbons
Results are presented to address three issues:
1.effect of changing ethane to propane ratio for fixed C2+
2.effect of changing C2+ for fixed propane to ethane ratio
3.design of a model fuel for testing purposes
We discuss only relative changes of fuel composition with
different C2+ content. The absolute errors in ignition time for
different reaction schemes are still considered to be too large
for gas turbine design purposes and subject to fundamental
research. (Goy et al. 2001). At ALSTOM, this issue has been
looked at in a co-operative research with DLR Stuttgart where
we have used shocktube experiments to calibrate and
benchmark different reaction schemes.
The relative difference between different reaction schemes
gives an indication of the uncertainty associated with such
Figure 5: Ignition times for variation in C2+ content.
Comparison of 3 different reaction mechanisms GRI, “A” and
“B”.
As shown in figure 5, propane and ethane addition to Swiss
Gas shortens auto ignition time for all mechanisms. The
relative decrease in ignition time is lower at high
concentrations of C2+. The mechanism differ by up to by 6%
C2+ for the same effect on ignition time or by up to 25% in
ignition time for constant C2+. This compares to a change of
up 25 K in operating conditions of the engine if the right
kinetical scheme is used.
2. To study the effect of the ethane to propane ratio for a
constant C2+value eight different fuel mixtures were tested: 92
% Swiss Gas + 0-8 % ethane + 8-0 % propane (C2+ = 10.71 % )
The graphs in figure 6 show that in mechanisms containing the
GRI chemistry ethane seems to react faster than propane while
the effect for the Wang chemistry is almost negligible at this
temperature. This difference is significant and has an impact for
example in the design of a model fuel where predicted ignition
properties can vary by up to 30%, depending on the choice of
the mechanism.

4 Copyright © #### by ASME

 1.3
Finally, it is mentioned that such chemical-kinetic analysis does
GRI3.0 A B not consider any effects which are specific to the combustor
1.2
design and which are related to turbulence-chemistry
1.1
natural gases interaction. The ignition process is strongly affected by the
relative igntion time

heat-up process of the fuel as it is injected into the highly

1.0
turbulent vitiated air stream. An appropriate model to describe
0.9
Birr Test gases
this process has been described in Brandt et al (2003) and
0.8 validated in Ivancic et al. (2004).
0.7

0.6 3. RIG & ENGINE TESTING

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
propane/(propane+ethane)
100% ethane 100% propane Atmospheric & high pressure rig tests
Figure 6: Relative ignition times for varying C2/C3 ratio at Mixing & combustion performance are initially tested in water-
constant C2+ (10.7%) with different mechanismus To study the channel and atmospheric combustion labs. Due to the strong
effect of the ethane to propane ratio for a constant C2+value pressure dependence of the combustion behavior these rigs
eight different fuel mixtures were tested: 92 % Swiss Gas + 0-8 have been mainly used for generic testing and to calibrate
% ethane + 8-0 % propane (C2+ = 10.71 % ) numerical design tools, and their use has been described for
example in Ivancic et al (2004).
3. Theoretical autoignition analysis has been used to “design” Full engine pressure tests with the model gases as
model fuels for rig and engine testing. The rational has been described in the previous section have been performed in
that key indicators for the model fuel are the Wobbe index, ALSTOM’s combustor test rig at DLR, Cologne, Germany. The
heating value, and ignition time based on an “equivalent C2+”. test rig contains both EV and SEV combustors at engine
The matching condition was primarily based on ignition time. pressures, temperatures and load conditions. Figure 7 shows the
The model gases for the engine tests contain Swiss natural gas, installation.
Propane and N2. The target were gas mixtures, which simulate
the customer gases with 10, 13.5 and 15% C2+ and inert
content up to about 19%. Fuel gases with normal inert content
but high C2+ content were looked at as well.
Matching the auto-ignition delay times between customer
gas and model gas means that the C2+ content may differ. At
the same time we had to account for the uncertainty mentioned
above (choice of kinetic scheme etc.). That bears some risk for
the tests. Mitigations actions such as special protection
measures had to be prepared for and were included in our test
schedule.
The inert content was specified to achieve the desired
heating value. As CO2 had to be substituted through N2 we had
to compromise in density for some of the mixtures. The Wobbe
index was allowed to differ as well.
The resulting model gases used in the GT26B engine tests Figure 7: ALSTOM’s combustor test rig at DLR, Cologne,
in Switzerland are listed in table 2 below. Germany.
CM Mix 1, 2, 3 Swiss Gas Thaigas Mix 1
Run 411 412 424 412 413 403 406 407 407 415 418 426 The tests were carried out with a variety of gas mixtures.
LHV
MJ/kg 46.4 46.5 46.4 46.5 46.5 47.4 47.3 47.2 32.5 33.3 32.3 34.0 One goal was to evaluate the consequences of replacing the
density
kg/m3 0.85 0.98 1.01 1.15 1.16 0.70 0.70 0.70 0.94 0.92 0.92 0.90 various higher hydrocarbons with Propane and replacing CO2
LHV
MJ/m3 39.6 45.6 46.9 53.4 53.7 33.2 33.3 33.2 30.4 30.5 29.6 30.7 with N2. The second goal was the definition of a preliminary
W .I.
MJ/m3 46.5 50.0 50.6 54.1 54.2 43.0 43.1 43.0 34.1 34.6 33.6 35.1 operating concept.
Actual C2+
% 16.9 28.0 30.9 42.7 43.1 3.8 4.5 4.3 13.8 13.0 12.4 12.6
With the flame sensing and monitoring equipment we were
Eq. C2+
% 10.4 13.3 14.1 16.0 16.3 3.7 4.2 4.1 10.2 9.9 9.8 9.7
able to show 1st that the EV 17i dry low-NOx burner is
Table 2: typical engine gas compositions with high C2+ content insensitive to high C2+ content and bears no flash back risk.
(CM), the standard natural gas at the Birr testing facility (Swiss 2nd, we have derived an operating concept for the SEV burner
gas), and gas mixtures for testing (Thaigas mix) based on Swiss depending on the C2+ content. With both findings we could
gas and nitrogene and propane addition. proceed to ALSTOM’s GT26B test facility in Birr, Switzerland.

5 Copyright © #### by ASME

Engine tests at ALSTOM’s GT26B test center in Birr
Based on the availability of Propane and Nitrogen in
sufficient quantities for a GT26B engine test a specialized fuel
gas mixing plant was designed, commissioned and operated at
the ALSTOM Test Center in Birr, Switzerland. Figure 8 shows
a sketch of the system layout and the plant overview containing
liquid propane and liquid nitrogen supplies, vaporizers and a
mixing station, which merges C3H8 and N2 with Swiss natural
gas from the grid. The installation of such a gas mixing facility
for a GT26 gas turbine operating at full load implied train loads
of propane and nitrogen, massive sprinkler systems, special
sealing of all basement corridors, tunnels and drainages, to
name but a few of the efforts which have been carried out for
validation.
The pulsation amplitudes of the SEV combustor are
recorded as well. We find that the delta in SEV burner inlet
temperature between the pulsations and the critical metal
temperature is independent of the C2+ content. That means that
the flame shape and relative position, which are described
through its characteristic stability remains unaffected by C2+.
The change of the flame position can therefore be controlled
only with the burner inlet temperature, just one single
parameter. This is a key result as it comes to the control effort
for handling high C2+ gases. It simplifies our operating
concept.

Layout

Steam
Generator Hall 11

Heater &
Pre-
Vaporisers
Heater
Room
Mixer
New Pipe

Nitrogen Storage
& Pump
Flare

Propane Storage
& Pump
Figure 9: Instrumented SEV burners with thermocouples for
Figure 8: Installation overview of the gas mixing plant at the flashback monitoring.
ALSTOM GT26B test center in Birr, Switzerland.

The gas mixing plant allows to vary the inert and C2+
content on-line, i.e. while the GT26B is running. Typically, the
engine is started on normal Swiss natural gas; then nitrogen is
added to the desired flow rate and finally propane is brought in
to the desired value. Then the test period follows. In reverse
order the composition is reduced back to Swiss natural gas. A
typical test period contains the extended mappings of both
combustors at different loads.
Acceptance criteria have been derived from material
temperature data at the SEV burner. That instrumentation is
shown in figure 9. One recognizes the thermocouples at the exit
of the mixing zone, which detect the approaching flame when Figure 10: SEV burner metal temperature for variations in
the C2+ content increases. We record a linear rise of the metal burner inlet temperature and fuel composition. General trends
temperature with the rise of the burner inlet temperature up to are indicated by straight lines.
the point when the flame reaches the vicinity of burner exit.
Then the rise becomes larger than linear. We set the limit for a Finally, we show how the results obtained at the single burner
sustainable operation at the maximal metal temperature allowed level have been confirmed by the engine tests. In figure 11 the
by the burner cooling design. That level of critical metal excellent agreement of the trend for the maximal allowable
temperature is reached for high C2+ content at a lower burner burner inlet temperature, for variations in C2+ addition is
inlet temperature than for low C2+ content, and this behavior is shown. Also included in the figure are the results from the first
schematically shown in figure 10. The flame movement is fleet engine where also detailed instrumentation was available.
monitored with cameras at the same time.

6 Copyright © #### by ASME


Figure 11: comparison of different rig and engine results for the
critical burner inlet temperature where the burner metal
temperature reaches its maximal allowable limit.
For completeness, it is noted that the EV combustor did not
show any flashback risk with increasing C2+ content with the
ALSTOM dry-low NOx burner EV17i which is installed in the
first combustor stage of the GT24B and GT26B engines. The
flame proximity has been monitored with burner shell and
lance tip temperature measurements. It has been observed that
the flame position does not change with the C2+ content in the
EV 17i burner and the flashback risk is negligible; in fact, the
only observable change has been the increased stability for lean
blow out with increased C2+ content which is even beneficial
for operation at low EV combustor temperatures.
We conclude that the two stage combustor system of
ALSTOM’s GT26B and GT24B with the EV and SEV burner
design is very well capable of handling varying C2+ contents
through a simple adjustment of the SEV burner inlet
temperature. Based on these findings, the operating concept
for the GT24B and GT26B engines has been modified to cover
low and high C2+ gases and is summarized in the diagram in
figure 12. There is no need for action up to 9% C2+ content.
Above 9% a moderate (and linear) reduction of the SEV burner
inlet temperature is applied up to the maximal allowable limit
which is given by today’s gas experience at 16%.
Also shown in the figure is the EV lean extinction
temperature which decreases for increasing C2+. As a
consequence, the operating window for the engine remains
wide for flexible operation for all gas compositions.
SEV Burner Inlet Temperature
SEV burner protection limit
Operation Window
EV burner lean extinction limit
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
C2+ [%]
Figure 12: The operation concept of the GT24B and GT26B
engines for low and high C2+ gases. Shown is only the range
of 9-16% C2+ where the operating concept requires a small
change in burner inlet temperature. Below 9% there is no
change in operating concept required.
4. FLEET IMPLEMENTATION
Fast gas composition monitoring
With the operation concept for varying C2+ content defined we
had one more task left to do. That was to detect and measure
C2+ containing gases fast enough.
For ALSTOM’s fuel flexibility test campaign in Birr,
Switzerland we had developed a gas sensor which records and
transmits any changes of the gas composition at the gas mixing
plant faster than the gas reaches the fuel distribution system of
the engine. For this application it was clear that a gas
chromatograph would be too slow for this task. A novel IR gas
sensor was developed, tested and commissioned and it
performed very well during the test.
For the field application we had received data from our
customers, which illustrate that the changes in gas composition
are sufficiently slow for gas chromatographs to record them.
All data had been taken with gas chromatographs however and
we have installed a fast gas sensor during the first C2+
implementation at the Bowin power plant as a precautionary
measure. We have learnt on that site however that even in large
gas networks the changes in gas composition can be much
faster than it is recorded by gas chromatographs. In one
example the trip of several stages of a gas separation plant
occurred. It caused a rise of the C2+ content from 10 to 16%
within 30 seconds at the GT. That separation plant is located
90km upstream of the gas turbine requiring a flow time of
45min. An example of the changes in gas composition are
illustrated in figure 13. One compares the readings from the
fast gas sensor (green) to the gaschromatograph data (blue).
7 Copyright © #### by ASME
 range of 10-16Vol% and the emissions are close to single digit
values.

100

GT NOx @15%O2, dry [ppm]

10

Guarantee
Eninge
GT26B2.1
A Ringsend 11
Eninge
GT26B2.0
B San Roque 21
1

0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Relative Load [%]

Figure 13: Example of a rapid change of gas composition and Figure 14: Typical NOx emission levels observed in customer
the corresponding engine adjustment of SEV burner inlet engines with high C2+ gases. The emission levels are well
temperature. One compares C2+ readings from the fast gas below the guarantee level for the entire load range.
sensor (green) and the gaschromatograph (blue).

Fleet implementation and experience CONCLUSION

The fleet implementation of ALSTOM’s high C2+ combustion
technology started with an extended mapping at the Bowin
power plant in Thailand. We had installed extensive
instrumentation, similar to the GT26B at the ALSTOM test
center in Birr, Switzerland. The commissioning proceeded well
while we have obtained valuable experience about the full C2+
range through trips of a gas separation plant which occurred
during the commissioning. Extra inspections and reduced
inspection intervals were applied to the SEV burner and SEV
lance during the first 6 months of operation. The purpose of the
inspections was to check that no overheating damage due to
flashback events had occurred. The results of these tests
showed not only that no damage had occurred but that the
condition of the SEV burners was in line with the experience of
engines running on low C2+ fuel gas. The high C2+ engines
then adopted the standard fleet inspection intervals. The fleet
implementation so far covers 5 GT24B’s and 13 GT26B’s with
a total of 200000 OH. 145000 OH have been operated at C2+
levels above 9%. There is no C2+ related problem or finding in
particular there is no flashback or any other damage to any
combustor hardware nor any loss of availability through C2+.
For example, two field engines were equipped with SEV burner
exit thermocouples similar to the Birr test engine. These
thermocouples were connected to ALSTOM’s monitoring and
diagnostic system. The data has indicated that no flashback
events have taken place in these engines. This fact has also
been confirmed by boroscope inspections. The availability of
C2+ engines is high. For example in Bowin the availability in
commercial operation lies above 97%.
Finally, the GT24/26 engines are running with best-in-class
emission levels also for high C2+ gases. In figure 14, we give
for examples from the fleet where the gas composition is in the
ALSTOM’s GT24B / GT26B engines have demonstrated their
capability to burn low and high C2+ fuels and cope with fast
changes of C2+ content. That fuel flexibility combined with
reliable operation and high availability at low emissions comes
as a benefit to ALSTOM’s customers.
This paper has outlined both the careful validation procedure,
in particular for the reheat combustion system, as well as
describing the operating fleet experience. Today ALSTOM’s
experience covers 18 engines in the GT24B/GT26B fleet with
high C2+ gas supplies. Another GT26B will be commissioned
on high C2+ gas this year, three more in 2005. These engines
have logged more than 200000 OH so far, 145000 of which are
reliable operation with high C2+ gases, i.e. with a C2+ content
above 9%.
REFERENCES
M. Brandt, W. Polifke, B. Ivancic, P. Flohr, B. Paikert
„Autoignition in a gas turbine burner at elevated temperature“,
ASME GT2003-38224
M. Braun-Unkhoff & C. Naumann, DLR, private communi-
cation
C.J. Goy, A.J. Moran, G.O. Thomas “Autoignition
characteristics of gaseous fuels at representative Gas turbine
conditions” ASME 2001-51.
GRI3.0: http://www.me.berkeley.edu./gri_mech.
F. Joos, P. Brunner, B. Schulte-Wenning, K. Syed
“Development of the sequential combustion system for the
ABB GT24/26 gas turbine family, ASME 1996-GT-315
Z. Qin, V. Lissianski, H. Yang, W. C. Gardiner, S. Davis, H.
Wang, 28th Proc. Comb. Symp. 1663 (2000).

8 Copyright © #### by ASME

View publication stats