Downloaded from https://iranpaper.

ir
https://www.tarjomano.com https://www.tarjomano.com

Proceedings of GT2006
ASME Turbo Expo 2006: Power for Land
Proceedings ofSea and Air
GT2006
May 8-11, 2006, Barcelona, Spain
ASME Turbo Expo 2006: Power for Land, Sea and Air
May 8-11, 2006, Barcelona, Spain

GT2006-90150

GT2006-90150

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

FULL SCALE TESTING OF A LOW SWIRL FUEL INJECTOR CONCEPT FOR
ULTRA-LOW NOX GAS TURBINE COMBUSTION SYSTEMS

1 2
Waseem Nazeer , Kenneth Smith, and Patrick Sheppard Robert Cheng and David Littlejohn
Solar Turbines Incorporated Lawrence Berkeley National Laboratory
2200 Pacific Highway One Cyclotron Road
San Diego, California 92101 Berkeley, California 94720
USA USA

ABSTRACT acceptable levels. This low emissions operability window is

The continued development of a low swirl injector for close to the lean extinction limit which can lead to high
ultra-low NOx gas turbine applications is described. An injector amplitude combustion pressure oscillations that may damage
prototype for natural gas operation has been designed, the combustor hardware.
fabricated and tested. The target application is an annular gas Lean premixed combustion systems first entered the gas
turbine combustion system requiring twelve injectors. turbine market in the early 1990’s. Typically, NOx levels were
High pressure rig test results for a single injector prototype reduced to 25 to 42 ppm range on natural gas fuel. As air
are presented. On natural gas, ultra-low NOx emissions were quality regulations continue to tighten, manufacturers are
achieved along with low CO. A turndown of approximately 100 actively pursuing technologies that can meet customer demands
0
F in flame temperature was possible before CO emissions of the future. In some parts of the U. S., gas turbines must
increased significantly. already meet regulations as low as 2.5 ppm NOx. To date lean-
Subsequently, a set of injectors was evaluated at premixed combustion systems have not been able to achieve
atmospheric pressure using a production annular combustor. these ultra-low levels. Operators must rely on costly exhaust
Rig testing again demonstrated the ultra-low NOx capability of clean up technologies such as selective catalytic reduction
the injectors on natural gas. (SCR).
An engine test of the injectors will be required to establish This paper presents the most recent results from a
the transient performance of the combustion system and to collaborative effort between Lawrence Berkeley National
assess any combustor pressure oscillation issues. Laboratory (LBNL) and Solar Turbines that is being supported
by the U.S. Department of Energy. The project is focused on
INTRODUCTION the assessment of a low swirl injector concept (LSI) conceived
Industrial gas turbine manufacturers rely primarily on lean- at LBNL. The goal of this work is to establish the capability of
premixed (LP) combustion technology to meet current engine the LSI technology to meet the need for ultra-low NOx
emissions regulations. LP combustion systems premix fuel and emissions in mid-range industrial gas turbines without SCR.
air prior to injection to prevent stoichiometric burning locally
within the flame. This results in a more spatially uniform flame DEVELOPMENT STRATEGY
temperature that reduces the conversion of atmospheric nitrogen The evolution of the LSI for gas turbines has followed the
to NOx. However, LP combustion does have shortcomings. By path shown in Figure 1. The development activity began with
operating at lean conditions, there is a tendency for increased the design of a proof-of-concept injector for Solar’s Taurus 70
carbon monoxide (CO) and unburned hydrocarbon emissions. (T70) gas turbine (7.5 Mw). The proof-of concept hardware
Consequently, the combustion system must operate within a was based on the initial concept evaluations completed at LBNL
narrow flame temperature range to keep all three species at [1]. Design evolution to a prototype status involved single

1
Corresponding author
2
Currently at Vanderbilt University

1 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

injector testing using an eight-inch diameter, back-side-cooled, HARDWARE CONFIGURATION

combustor can at both atmospheric and high pressure (190 A cross-section of the prototype injector is shown in Figure
psia). 2. The injector has two concentric, axial flow passages. Each
The work has progressed to a full-scale, atmospheric flow passage is fed with natural gas and provides the residence
pressure rig test of a set of twelve T70 injectors in a production time needed for good air/fuel mixing. The outer flow passes
T70 annular liner. Future efforts will include T70 engine tests through an axial air swirler (vane angle of approximately 45
to evaluate transient performance and durability. degrees). The inner flow passes through a perforated plate and,
The results reported here document the first test of multiple thus, has no swirl component. The typical swirl numbers for the
LSI injectors in an annular combustor configuration. LSI are between 0.4-0.5. The open area of the perforated plate

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

is used to fine-tune the air split between the swirled and
unswirled flows.
Natural gas is delivered to the injector flow passages
through a series of radial fuel spokes. Each spoke contains a
Design and
series of injection orifices that feed both the inner and outer air
Prototype
Fabrication flows. The design intent is to maintain both the swirled and
unswirled flows at the same nominal equivalence ratio.
Both air/fuel streams mix in a post-mix zone at the
downstream end of the injector. The flame that is produced is
aerodynamically stabilized just downstream of the injector exit
Single Injector plane. A pilot fuel injector is located at the center of the
Low Pressure injector for added flame stability during transients and off-
Evaluation design operating conditions.
More complete descriptions of the LSI concept and early
Optimization development work have been presented elsewhere [1, 2, 3].
Design
No
Criteria Met?

Yes Fuel Manifold

Single Injector
High Pressure Swirl Vanes
Evaluation

Design
No
Criteria Met?

Yes Pilot
Full Annular
Atmospheric
Perforated Plate
Test No

Figure 2. Cross Section of the Low Swirl Injector

Design
Criteria Met?
EXPERIMENTAL FACILITY
Three different test facilities are being used to support the
yes
injector development. These include:
Engine Test
- A single injector, atmospheric pressure rig (non-
combusting) to measure fuel-air ratio profiles at the
injector exit plane. The facility allows measuring both
radial and circumferential profiles of methane
concentration.
Figure 1. Injector Development Strategy
- A high-pressure, single-injector test rig with a 9 pps,

2 Copyright © 2006 by ASME

Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

300 psi air flow capability (Fig. 3). An indirect air All of the combustion facilities include instrumentation to
heater provides combustor inlet temperatures up to 800 monitor fuel and air flows, system pressures (static and
0
F, the T70 full load design point. dynamic), temperatures, and emissions (NOx, CO, CO2, O2,
UHC).
All testing is being done using commercial grade natural
gas. A typical gas composition is shown in Table 1.

Component Percent

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

Methane 95.83
Ethane 1.78
Propane 0.31
I-butane 0.04
N-butane 0.08
I-pentane 0.02
N-pentane 0.02
Carbon dioxide 0.85
N-hexane 0.03
Nitrogen 1.04

Table 1: Typical Natural Gas Composition

Figure 3. High-Pressure Single Injector Test Rig

LOW PRESSURE SINGLE INJECTOR TESTS

- An atmospheric pressure rig (Fig. 4) capable of
providing 10 pps air flow capability at 1200 0F. This Fuel-Air Mixing Effectiveness
rig is being used to test the prototype, multi-injector, One of the key factors affecting NOx emissions is fuel-air
annular combustion system. homogeneity at the injector exit. Unmixedness can cause
regions of high equivalence ratio where the local flame
temperature is higher than the average. This can lead to high
NOx [4]. In addition, an ineffective premixer can cause auto-
ignition/flashback, particularly in high pressure ratio,
recuperated engines.
Premixing tests were performed at atmospheric pressure
and ambient temperature. Air flow was set to match residence
times within the injector (Table 2). A sampling rake having
eight separate, equally–spaced, radial sampling tubes was used.
The rake was rotated in 5-degree increments to cover the entire
injector exit plane. The mixture from each sample tube was
analyzed with a non-dispersive infra-red (NDIR) methane
analyzer.
A color contour plot of the full load premixing data is
shown in Figure 5. The data were normalized using the grand
average and are shown as percentages of the average. Figure 5
shows that the radial non-uniformity is larger than the
circumferential non-uniformity. Radial non-uniformity is
approximately +/- 20% while the circumferential variation is
near +/- 5%. The relatively large radial non-uniformity is due
primarily to the test having been run without fueling of the pilot
Figure 4. Atmospheric Pressure Multi-Injector Test injector.
Rig

3 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

by maintaining the inlet air temperature and air flow constant

Pressure Drop (%) 4 while reducing the fuel flow to the injector. The fuel was
Air-to-fuel Ratio 30 slowly reduced until a flame out occurred. The results from
Intervals (degrees) 5 several tests at different air flows are shown in Figure 6.
Figure 6 shows LBO to occur at flame temperatures near
Pilot (%) 0
2500 0F, essentially independent of the injector exit velocity.
The lifted LSI flame (aerodynamic stabilization) maintains the
Table 2: Test Conditions for Fuel-Air Mixing Test same location relative to the injector exit plane despite changing
velocities. This behavior has been reported earlier [1]. The

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

insensitivity to injector exit velocity suggests good performance
at transient conditions, although engine testing is needed to
verify this assumption.

2800

Adiabatic Flame Temperature (0F)

2700

2600

2500

2400

2300

2200
2100

2000
150 170 190 210 230 250 270
Velocity (Ft/Sec.)

Figure 6. Lean Blow Off Atmospheric Tests

Figure 5. Fuel-Air Mixing Uniformity

Figure 5 shows that the left half of the injector is somewhat High Pressure Single Injector Test
richer than the right half. This has been attributed to air
blockage by the pilot fuel tube that extends across the injector Emissions Performance
inlet. The high-pressure single injector rig tests were conducted
Although it was clear that the injector fuel-air homogeneity using a single 8 in. diameter can combustor. The combustor
could be improved further, earlier high pressure testing had was sized to match the residence time of the full scale, annular
shown excellent emissions performance. Thus the decision was T70 combustor. Testing was conducted at simulated part and
made to proceed with the current design to multi-injector full load operating points of the T70 (Table 3). The purpose of
testing. The intent was to assess other combustion system these tests was to quantify emissions, lean stability, and injector
performance factors such as transient response and combustor temperatures at actual engine conditions.
pressure oscillations before focusing on improvements in The rig tests were run by fixing the combustor inlet air
premixing effectiveness. flow, pressure and temperature. The total fuel was then varied
in increments to generate emissions maps as a function of
Lean Blow Off (LBO) Tests primary zone equivalence ratio or adiabatic flame temperature.
LBO data were collected as part of a study to establish the The NOx, CO, and adiabatic flame temperature data has
optimum design point equivalence ratio for the LSI. Tests were been normalized with the targeted design points of 5 ppm, 25
run using an eight-inch diameter by eleven-inch long quartz ppm, and 2700 0F, respectively. Full load NOx and CO data
tube to simulate the geometry of the high pressure, single were collected at different pilot fueling rates and are shown in
injector combustor liner. A single LBO data point was obtained Figure 7.

4 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

Parameter Part Full As expected, flame stability improved with increasing pilot
Load Load fuel flow rates, although at the cost of higher NOx emissions.
Combustor Inlet Temp. (F) 750 800 A second test was conducted for simulated part load
Combustor Inlet Pressure (PSIA) 153 195 conditions (Fig. 8). In these tests the combustor inlet air
Pressure Drop (%) 4 4 temperature and pressure were reduced while maintaining the
Adiabatic Flame Temps (F) 2600 to 2600 to pressure drop across the injector. The NOx and CO data
3000 3000 showed similar behavior to the full load emissions. Reducing
Pilot (%) 0, 5, 10 0, 5, 10 the combustor inlet temperature and pressure did not show a
significant impact on NOx and flame stability.

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

Table 3: Single Injector High Pressure Test Conditions

Figure 7 shows that with low pilot fueling levels, a

reasonable operating range was achieved with ultra-low NOx 10
9 0 % Pilot

Normalized NOx @ 15 % O2
emissions. Testing showed single digit NOx and CO emissions
5 % Pilot
with 5% pilot fueling with a 100 0F turndown margin in flame 8
10 % Pilot
temperature. The 0 % and 5 % pilot data showed no significant 7
difference in NOx emissions although CO was higher for 0% 6
pilot at the leanest condition. 5
4
3
10
2
9 0 % Pilot 1
Normalized NOx @ 15 % O2

5 % Pilot 0
8
10 % Pilot
7 3.5
Normalized CO @ 15 % O2 (PPM)

0 % Pilot
6 3 5 % Pilot
5 10 % Pilot
4 2.5
3 2
2
1 1.5
0 1
3.5 0.5
3 0 % Pilot 0
Normalized CO @ 15 % O 2

0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1

2.5 5 % Pilot
10 % Pilot Normalized Adiabatic Flame Temperature
2
Figure 8. High Pressure NOx and CO Emissions at
1.5 Simulated Part Load
1
0.5 Thermal Paint Test
Injector material temperatures were assessed using
0
0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1 temperature-sensitive (thermal) paint. An injector was coated
with thermal paint (TP8) and run in the high pressure rig at full
Normalized Adiabatic Flame Temperature load conditions (Table 4).
Results from the test are shown in Figure 9. Post-test
Figure 7. High Pressure NOx and CO Emissions at analysis showed that the majority of the main barrel and the
Simulated Full Load central perforated plate remained below 915 0F. These low
temperatures are attributable to the lifted form of the LSI flame
The 10 % pilot data showed significantly higher NOx and a and represent one of the major benefits of the LSI concept [1].
more rapid rise in NOx emissions with increasing flame The exit portion of the injector barrel experienced slightly
temperature. This is likely due to the fuel rich condition that higher temperatures. The outer and inner surfaces of the injector
exists along the injector axis at 10% pilot fuel flow. barrel near the injector exit plane experienced temperatures

5 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

between 915 0F and 1350 0F. This disparity in the After successful completion of the single injector high pressure
circumferential temperature distribution is attributable to tests, a full set of injectors was fabricated and tested in a full
variable air leakage between the injector body and a sealing scale, T70 annular combustor liner. The purpose of the test was
grommet that is part of the combustor liner. Inspection of the to evaluate potential flame interactions and to document
injector after testing revealed that paint at the hottest area combustor exit temperature profile, pattern factor, lean blow off
showed signs of rubbing by the grommet, suggesting a tight seal equivalence ratio and emissions.
in this location. Locations where the seal was not as tight
experienced more air leakage which cooled the injector barrel Combustor Exit Temperature Profile
locally. In any event, the injector temperatures were well within Once again, simulated part and full load conditions of a

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

the general guidelines for meeting service life requirements. T70 engine were run (Table 5). The combustor exit
temperature profile was measured using four thermocouple rake
assemblies mounted at the exit plane of the combustor (90
degrees apart). These rakes were rotated circumferentially
Parameter Condition
around the combustor axis to provide a complete exit plane
Thermal Paint Type TP-8
temperature map. Each rake assembly consisted of seven
Combustor Inlet Temperature (0F) 800 radially-spaced thermocouples that spanned the annulus
Combustor Inlet Pressure (PSIA) 190 between the outer and inner combustor liners.
Pressure Drop (%) 4
Flame Temperature (0F) 2800 Parameter Part Load Full Load
Test Duration (Minutes) 30 Combustor Inlet Temp. (0F) 750 800
Combustor Air Flow (PPS) 3.7 2.9
Table 4: Single Injector High Pressure Thermal Paint Test Adiabatic Flame Temps (0F) 2600 to 3000 2600 to 3000
conditions Pilot (%) 5, 30 0, 5, 10

Table 5: Conditions for Multi Injector Atmospheric Test

Full load tests were run with pilot fueling levels of 0, 5, and
10%. The exit temperature profiles were all very similar.
Figure 10 shows typical results at full load and 10 % pilot fuel.
In Fig. 10, the temperature data have been normalized with the
bulk average and then plotted as a contour map.
The exit plane temperature measurements indicated good
circumferential uniformity and an appropriate radial
temperature distribution with relatively cooler regions close to
the inner and outer liners. The pattern factor was 0.18, well
within acceptable limits for the T70 engine. The exit
temperature profile and pattern factor were deemed acceptable
to ensure good hot section durability in an engine.

Figure 9. High Pressure Thermal Paint Test

FULL SCALE ANNULAR ATMOSPHERIC TEST

6 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

1.2
0 % Pilot

Normalized NOx @ 15 % O2
1
5 % Pilot
10 % Pilot
0.8

0.6

0.4

0.2

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

0
2.5
0 % Pilot
2 5 % Pilot

Normalized CO @ 15 % O 2
10 % Pilot
1.5

0.5
Figure 10. Full Annular Combustor Exit Profile
0
Emissions and Lean Blow Off Test 1 1.02 1.04 1.06 1.08 1.1 1.12
Emissions data were also collected for the full annular
Normalized Adiabatic Flame Temperature
combustion system at simulated T70 part load and full load
conditions. The emissions samples were collected using an air- Figure 11. Full Annular Combustor NOx and CO
cooled probe at one location at the combustor exit plane. Emissions at Simulated Full Load at Atmospheric
The NOx and CO data for full load conditions are shown in Conditions
Figure 11. The data for 0, 5, and 10 % pilot fuel flows are
presented. The emissions trends from the full scale atmospheric LBO tests were conducted with the full annular combustor
tests were very similar to the single injector high pressure tests. liner to determine how increased pilot fuel rates enhance
However, the NOx emissions levels for the full system stability. The results are shown in Figure 13. As expected,
atmospheric test were consistently lower than single injector increases in the pilot fuel flow to extremely high levels (30 and
high pressure tests. This is particularly apparent at the higher 50%) significantly lowered the LBO flame temperature. This
flame temperatures and at higher pilot fuel flows. It is unclear enhanced stability can be exploited in developing the control
whether this difference is attributable to: the lower combustor strategy for the T70 during lightoff and during transient
pressure of the full system test; an effect of injector interactions; operation when temporary emissions spikes may be permitted.
or simply the result of taking only a single point emissions
measurement for the full scale test. It is known that the
reactions forming NOx in the fuel rich zones caused by higher
pilot flows are highly dependent on pressure [4, 5, 6]. Engine
testing of the combustion system will provide more insight into
the results.
Full annular part load emissions data were gathered at 5,
30, and 50 % pilot fuel flows (Fig. 12). Again the 5 % pilot
NOx data are lower than the single injector data at pressure.
Performance at the very highest pilot flows were run to explore
possible modes of combustor operation during engine light off.

7 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

20
18 5 % Pilot CONCLUSIONS
The performance of a prototoype low swirl injector for
Normalized NOx @ 15 % O2

16 30 % Pilot
ultra-low gas turbine emissions has been demonstrated
14 50 % Pilot
successfully on natural gas. Testing has included high pressure,
12 single injector rig tests and atmospheric pressure tests of a full
10 set of injectors in an annular combustor rig.
8 The single injector tests showed that NOx emissions for
6 pilot fuel flows near 5% were below 5 ppm with low CO
4 emissions. The emissions performance was similar at simulated

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

2
full and part load conditions which suggests a wide range of low
emissions capability on an engine. Thermal paint tests at high
0
pressure have shown that the injector temperatures were very
2
low, consistent with long service life.
1.8 5 % Pilot The atmospheric pressure tests, using twelve LSIs and an
Normalized CO @ 15 % O 2

1.6 30 % Pilot annular Taurus 70 combustor liner, demonstrated equally low

1.4 50 % Pilot emissions levels although the low pressure emissions data were
1.2 relatively limited.
1 Overall, the combustor ran very well showing excellent
0.8 light-around characteristics, no indication of combustion
harmonics, and acceptable pattern factor and radial exit
0.6
temperature profile.
0.4
The next step in the development of the LSI will be a short
0.2 duration engine test to validate system emissions further,
0 document liner wall temperatures, establish control strategies
1 1.02 1.04 1.06 1.08 1.1 1.12 for lightoff and load transients, and document any occurrences
of high magnitude combustion oscillations that might require
Normalized Adiabatic Flame Temperature
further injector optimization.

ACKNOWLEDGMENTS
Figure 12. Full Annular Combustor NOx and CO The authors would like to thank Steve Waslo and Debbie
Emissions at Simulated Part Load at Atmospheric Haught, Program Managers at the U. S. Department of Energy
Conditions for their continuing interest in this project.

REFERENCES
60
[1] M. R. Johnson, D. Littlejohn, W. A. Nazeer, K. O.
50 Smith, R. K. Cheng, “ A Comparison of Flow Fields and
Emissions of High-Swirl Injectors and Low-Swirl Injectors for
40 Lean Premixed Gas Turbines, “ Proceedings of the Combustion
Pilot (%)

30 Institute 30, 2867-2874, 2005.

20 [2] S. E. Fable and R. K. Cheng, “ Optimization of a

Premixed Low-Swirl Burner for Industrial Applications, “
10 Second Joint Meeting of the U. S. Sections of the Combustion
Institute, Paper # 109, 2001.
0
2100 2200 2300 2400 2500 2600 2700
[3] R. K. Cheng, D. Littlejohn, W. Nazeer and K. O. Smith,
Adiabatic Flame Temperature (0F) “Laboratory Studies of the Flow Field Characteristics of the
Figure 13. Full Annular Combustor Lean Blow Off Test Low-Swirl Injectors For Adaptation to Fuel Flexible Turbines,”
Simulated Part Load at Atmospheric Condition International Gas Turbine Institute, Paper # GT2006-90878.

[4] K. O. Smith, “NOx Reduction for Small Gas Turbine

Power Plants, “ EPRI AP-5347, Final Report, 1987.

8 Copyright © 2006 by ASME

 Downloaded from https://iranpaper.ir
https://www.tarjomano.com https://www.tarjomano.com

[5] Leonard, G. and Stegmaier, J., “Development of an

Aeroderivative Gas Turbine Dry Low Emissions Combustion
System,” Journal of Engineering for Gas Turbines and Power,
Vol. 116, pp. 542-546, 1993.

[6] Leonard, G. and Correa, S., “NOx Formation in Lean

Premixed Methane Flames,” Second ASME Fossil Fuel
Combustion Symposium, PD-30, pp. 69-74, 1990.

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT2006/42363/107/4564821/107_1.pdf by University of Edinburgh user on 12 February 2024

9 Copyright © 2006 by ASME