Experimental Investigation of a New Combustor Model for Gas Turbines
by
M.J. Melo(1), J.M.M. Sousa(1), M. Costa(1) and Y. Levy(2)
(1)
Instituto Superior Técnico, Mechanical Engineering Department, Av. Rovisco Pais, 1049 - 001 Lisboa, Portugal
E-Mail: mcosta@navier.ist.utl.pt
(2)
Faculty of Aerospace Engineering, Technion, Israel Institute of Technology, Haifa 32000, Israel
E-Mail: levyy@aerodyne.technion.ac.il
ABSTRACT
The study reported herein aims to provide information on the present performance of a new combustor model for gas
turbines both under non-reacting and reacting conditions. The flow characteristics of the combustor model under
non-reacting conditions have been investigated using Laser-Doppler Anemometry. Data is reported for mean and
turbulent velocities as a function of the air mass flow rate and its preheating temperature. The isothermal flow
characterization was followed by combustion measurements at the exit of the combustor model. Measurements of mean
gas species concentration (O2, CO2, CO, HC and NOx) are reported as a function of the equivalence ratio and thermal
input for two different configurations of the main air inlets. The isothermal data revealed that (i) a common feature to
all test conditions is the establishment of a large recirculation zone, as shown in Fig 1; (ii) mean and turbulent velocities
increase within the recirculation zone as the air mass flow rate increases and (iii) the effect of the air preheating
temperature on the flow field is marginal. Under reacting conditions the data revealed that (i) combustion performance
is higher for both lower values of thermal input and equivalence ratio; (ii) NOx emissions are very low regardless of the
combustor operating conditions and (iii) the effect of the air inlet configuration on both combustor performance and
pollutant emissions is marginal.
|V| [m/s]
18.0
Air
16.8
15.6
14.4
13.2
12.0
10.8
9.6
8.4
7.2
6.0
4.8
3.6
2.4
1.2
0.0
V [m/s]
Fig. 1. Typical result for the isothermal mean flow structure within the combustor model
1
1. INTRODUCTION
The increase, in recent years, in the efficiencies of Gas Turbines (GT) has occurred regardless of a tightening in
environmental legislation on the emission of nitrogen oxides (NOx). The hot primary zone within the combustor and the
higher turbine inlet temperatures, a prime factor for increased efficiency, tend to increase the formation of NOx. Dry
low-NOx combustion systems have therefore been developed, but it appears that there is still a great need for low-cost,
safe and reliable NOx reduction methods especially for high efficiency small GT engines.
The primary objective of the present study is to develop innovative combustors for small GT, suitable for safe and
reliable operation at high temperatures while maintaining low NOx levels (typically less than 20 ppm). As an example,
for a GT with conventional efficiencies of the turbine (87%) and compressor (85%), operating at about 15 bars, a
Turbine Inlet Temperature (TIT) above 1600 K is required to boost the thermal efficiency over the value of 35%. A
similar concept is applicable to larger GT incorporated in Combined Rankin–Brayton Cycle (GTCC) to exceed 60%
efficiencies of energy conversion. Hence, relatively high exit combustion temperatures are required for high efficiency
GT systems.
The low-NOx emission levels from flameless oxidation combustors enable the operation of GT at higher TIT levels (in
comparison to conventional dry low-NOx technologies) thus maintaining elevated efficiency values and by that
enabling reduction in CO2 emissions. In power systems which incorporates a heat exchanger (the regenerative cycle),
the effect of TIT is even more significant.
In a GT combustor, there is only a minor link between the maximum flame temperature within the combustor (the
major parameter which affects NOx production) and the TIT value. High combustion temperature values (in the primary
combustion zone) are typically used to improve flame stability. The maximum TIT values are determined by the ability
of the turbine blades to operate reliably while exposed to the high temperature gases and being heavily loaded by the
centrifugal stresses. Significant cooling using dilution air is applied to the combustion products of the primary zone in
order to reduce their temperature to allowable TIT values. Dry low-NOx techniques, as Lean Premixed Pre-Vaporizer
(LPP) or Catalytic Combustion (CC) are typically based on attempts to reduce the flame temperature of the primary
zone while maintaining flame stability. The method proposed in the present FLOXCOM program1, presents an
innovative alternative that is aimed for clean and efficient GT. It is based on a technologically combustion solution, the
flameless oxidation method. This promising technology allows operation of the combustor at high temperature with
ultra-low NOx levels. In addition, this combustion mode has further advantages over other advanced NOx reduction
technologies such as safety, reliability and the possibility for its incorporation in a heat exchanger cycle using high air
temperatures at the combustor inlet. Flameless oxidation requires exhaust gas recycling, whether it is internal or
external. It is an effective method to reduce flame temperature and thereby NOx emissions. Unfortunately, the
efficiency of this method is affected by the maximal possible amount of recirculated hot exhaust gas. Flame instabilities
and eventually blowout may occur if the burner is operated with too low internal recirculation rates of combustion
products. Whenever the temperature of the mixture of the recycled exhaust gas with the fresh air exceeds the autoignition temperature of the fuel, the fuel is ignited automatically and continuous combustion is sustained. In contrast to
classic diffusion flame, temperature peaks can be avoided even at high air preheat temperatures in this combustion
mode. It is characterized by a moderate and distributed temperature rise, small gradient of species or temperatures, low
radiation emissions and noise. Therefore, the thermal-NO formation can be largely suppressed. Other advantages are:
improved internal combustor aerodynamics for more uniform wall temperature and lower wall cooling requirements;
lower thermal stresses, and lower values of the exhaust gas pattern factors for more circumferential uniformity at the
combustor exit. Both of these advantages are relevant for increased Mean Time Between Failure (MTBF) and
reliability of the GT.
The main objective described in the present paper is directed to the development of an innovative combustor for GT. It
can also be applied to aero-engines for both helicopters and jet aircrafts. The application of the flameless oxidation
combustion concept to GT combustors and the role of the flameless oxidation mode in NOx emission reductions was
already confirmed by computer simulations (Levy et al., 2004a). Flameless oxidation is also being currently used in
industrial furnaces with regenerative burners in non-adiabatic cycle, indicating extremely low-NOx emission levels
(typically less than 5 ppm), much below the environmental requirements.
1
The FLOXCOM program is conducted through the support of the FP5 program of the EC (ENK5-CT-2000-00114, see http://floxcom.ippt.gov.pl)
by a consortium of industrial and academia partnership that consists of eight parties from different European Countries and Israel. The partners
include departments of universities and engineering firms having expertise in the development of GT engine evaluation, and an R&D center of an
industrial group.
2
The technological objectives achieved and described in the present paper are to test a sector of a pilot combustor using
the flameless oxidation combustion concept at atmospheric pressure. More specifically, it describes the effort invested
to improve the combustion air injection mode and distribution. This is done by numerical (elsewhere) and experimental
investigation of the flow field and combustion features. Different combustor models (of the 60-degree sectors) were
produced with optical windows, for detailed local measurements of the velocity field, temperatures and species
concentrations distributions under non-reactive and reactive conditions. These were used for further adjustments of the
different models and for comparisons with the CFD predictions (Levy and Arfi, 2004 and Levy et al., 2004a, 2004b).
2. EXPERIMENTAL METHOD and PROCEDURES
Figure 2 shows a schematic of the combustor cross section of the model. In this work, two different configurations for
the Air 1 (see Fig. 2) inlets were used, as shown in Figs. 3a) and 3b).
Air 1
Fuel
Air 2
Exhaust gases
Fig. 2. Schematic of the combustor model
Prior to the combustion experiments, the isothermal flow in the combustor model was thoroughly characterised by the
application of Laser-Doppler Anemometry (LDA), as shown in Fig. 4. A two-component velocimeter from DANTEC,
which was operated in the dual-beam backward-scatter mode, was employed to meet this objective, see, e.g., Melo et
al. (2002). High data rates, close to 1 kHz, were obtained by seeding the flow with small droplets of a mixture prepared
with ethyleneglycol (20%) and water (80%). The droplets were generated by four medical nebulisers INSPIRON
002305-A. The back-scattered light collected by the receiving optics was band-passed filtered and processed by two
DANTEC 57N20/57N35 Burst Spectrum Analyzers interfaced with a IBM AT compatible computer. Velocity statistics
were evaluated by ensemble averaging, calculated from 10,000 samples, using BURSTware software. Errors incurred
in the measurement of velocities by displacement and distortion of the measuring volume due to refraction on the
combustor model walls were found to be negligibly small and within the accuracy of the measuring equipment in the
case of the LDA.
3
(a)
(b)
Fig. 3. Schematics of the flow configurations for the Air 1 (see Fig. 1)
a)
Original Configuration
b)
Modified Model Configuration
Slave BSA
Master BSA
2x Oscilloscope
Photo multipliers
Laser (Ar+-ion)
Fiber optics
3-D traverse
system
Back-scattering probe
Seeding
Pressurized
air
Test rig
Pressurized air for
windows cleaning
Air 1
P
P
Fan
T
Flowmeter
T
Preheater
Pressurized Air 2
Fuel (CH4)
Fig. 4. LDA measurement system
Figure 5 shows the measurement system for the combustion tests. The gases for the measurement of the flue-gas data mean O2, CO, CO2, hydrocarbons (HC), and NOx concentrations - were withdrawn using a water-cooled stainless steel
probe. The wet sample was drawn through the probe and part of the sampling system by an oil-free diaphragm pump. A
condenser removed the main particulate burden and condensate. A filter and a drier removed any residual moisture and
4
particles so that a constant supply of clean dry combustion gases was delivered to each instrument through a manifold
to give species concentrations on a dry basis. The analytical instrumentation included a magnetic pressure analyzer for
O2 measurements, nondispersive infrared gas analyzers for CO2 and CO measurements, a flame ionization detector for
HC measurements, and a chemiluminescent analyzer for NOx measurements. The analog outputs of the analyzers were
transmitted via analogic/digital (A/D) boards to a computer where the signals were processed and the mean values
computed. Zero and span calibrations with standard mixtures were performed before and after each measurement
session. The maximum drift in the calibration was within ± 2% of the full scale. At the combustor exit, where the gas
composition was nearly uniform, probe effects were negligible and errors arose mainly from quenching of chemical
reactions and sample handling. Samples were quenched near the probe tip to about 150 °C and condensation of water
within the probe was avoided by controlling the inlet temperature of the cooling water (typically to around 60 °C).
Repeatability of the flue-gas data was, on average, within 5%.
12 bit A/D
FID analyser
(UHC)
Probe
88888
88888
Chemiluminescent
analyser (NOX)
Paramagnetic analyser
(O2)
88888
Infrared analyser
(CO, CO2)
Exhaust
Termocouple
type K
Cotton wool
Condenser
Zero Span
gas
gas
Silica gel
Water
Filter
Dryer
Water
Diaphragm
pump
Pressurized air for
windows cleaning
Test rig
P
Exhaustion
T
Air 1
Flowmeter
P
T
Pressurized
Fuel
Air 2
Preheater
Fan
Thin-plate orifice
(CH4)
Fig. 5. Combustion measurement system
3. RESULTS AND DISCUSSION
Table 1 summarises the operating conditions for the experiments under non-reacting conditions, which allow for an
assessment of the effects of the air mass flow rate (Air 1, see Fig. 2) and its preheating temperature on the flow
characteristics.
Table 1. Operating conditions for the experiments under non-reacting
conditions for the original model configuration*
Test condition
Air 1 flow rate (kg/s)
Air 1 temperature (K)
1
0.12
293
2
0.12
330
3
0.095
330
4
0.07
330
* For all conditions: Air 2 flow rate = 0 kg/s; Fuel (air) flow rate = 0.00317 kg/s;
Fuel (air) temperature = 293 K.
5
Figure 6 shows the effect of the flow rate of Air 1 on the mean flow structure. The figure reveals that a common feature
to all test conditions is the establishment of a large recirculation zone and that the mean velocities increase within the
recirculation zone as the Air 1 mass flow rate increases. The figure also reveals that near the combustor outlet the mean
velocities are higher for the intermediate Air 1 mass flow rate tested (test condition 3 in Table 1). Figure 7 shows the
effect of the flow rate of Air 1 on the turbulence kinetic energy contours. It is seen that they are consistent with the data
presented in Fig. 6. Figure 8 shows the effect of the Air 1 inlet preheating temperature on the mean flow structure. As it
can be seen, this has only a marginal effect on the flow field.
Test condition 2
Air 1: 0.12 kg/s
Test condition 3
Air 1: 0.095 kg/s
Test condition 4
Air 1: 0.07 kg/s
|V| [m/s]
|V| [m/s]
|V| [m/s]
18.0
18.0
16.8
16.8
15.6
15.6
14.4
14.4
13.2
13.2
12.0
12.0
10.8
10.8
9.6
9.6
8.4
8.4
7.2
7.2
6.0
6.0
4.8
4.8
3.6
3.6
2.4
2.4
1.2
1.2
0.0
0.0
V [m/s]
18.0
16.8
15.6
14.4
13.2
12.0
10.8
9.6
8.4
7.2
6.0
4.8
3.6
2.4
1.2
0.0
V [m/s]
V [m/s]
Fig. 6. Effect of the Air 1 flow rate on the mean flow structure
Test condition 3
Air 1: 0.095 kg/s
Test condition 4
Air 1: 0.07 kg/s
Test condition 2
Air 1: 0.12 kg/s
k [m2 /s2]
k [m2/s2]
k [m2/s2]
19.0
19.0
19.0
17.8
17.8
16.6
16.6
15.4
15.4
14.2
14.2
13.0
13.0
11.8
11.8
10.6
10.6
9.4
9.4
8.2
8.2
7.0
7.0
5.8
5.8
4.6
4.6
3.4
3.4
2.2
2.2
1.0
1.0
17.8
16.6
15.4
14.2
13.0
11.8
10.6
9.4
8.2
7.0
5.8
4.6
3.4
2.2
1.0
Fig. 7. Effect of the Air 1 flow rate on the turbulence kinetic energy contours
Test condition 1
Temperature of Air 1: 293 K
Test condition 2
Temperature of Air 1: 330 K
|V| [m/s]
18.0
16.8
15.6
18.0
16.8
15.6
14.4
14.4
13.2
13.2
12.0
12.0
10.8
10.8
9.6
9.6
8.4
8.4
7.2
7.2
6.0
6.0
4.8
4.8
3.6
3.6
2.4
2.4
1.2
0.0
V [m/s]
Fig. 8. Effect of the Air 1 inlet temperature on the mean flow structure
6
|V| [m/s]
1.2
0.0
V [m/s]
Table 2 summarises the operating conditions for the experiments under reacting conditions, which allow for an
assessment of the effects of the equivalence ratio and thermal input on pollutant emissions for both configurations
studied (see Figs. 3b and 3c).
Table 2. Operating conditions for the experiments under reacting conditions for both model configurations
Air 1 flow rate
Air 1
Equivalence ratio ( φ )
Methane flow rate (kg/s)
Model
(kg/s)
temperature (K)
-5
-4
Original (Fig. 2a)
0.012-0.038
361-387
7.8x10 -4.2x10
0.013-0.34
Modified (Fig. 2b)
0.012-0.035
378-417
1.2x10-4-5.1x10-4
0.13-0.49
Figure 9 show the measured flue-gas data for the original model configuration, shown in Fig. 3a, as a function of the
equivalence ratio and of the thermal input. The figure reveals that combustion performance is higher for both lower
values of thermal input and equivalence ratio. It is interesting to note that NOx emissions are very low regardless of the
combustor operating conditions.
Total thermal input (kW)
12.0
7.8
12.0
10000
22
9000
20
10000
CO2
18
CO
CO
NOx
16
HC
7000
14
6000
12
5000
10
4000
8
3000
6
4
2000
2
1000
0.10 0.15 0.20
7.8
12.0
0.25 0.30 0.35
0.40
0.45
0
0.50
HC
14
5000
4000
8
3000
6
4
2000
2
1000
0.10 0.15 0.20
0.25 0.30
0.35 0.40
Total thermal input (kW)
16.3
7.8 12.0
CO2
18
CO
NOx
14
10000
22
9000
20
8000
7000
6000
12
5000
10
4000
8
3000
6
4
2000
2
1000
0.10 0.15 0.20
0.25 0.30 0.35
0.40
0.45
16.3
0
0.45 0.50
20.9
10000
O2
20
HC
7000
10
Total thermal input (kW)
16
8000
6000
Equivalence ratio
O2
9000
12
Equivalence ratio
22
0
0.00 0.05
NOx
16
0
0.00 0.05
CO2
CO
18
O2, CO 2 (dry volume %)
0
0.00 0.05
O2, CO 2 (dry volume %)
8000
O2, CO 2 (dry volume %)
CO2
18
CO, HC, 100 NO x (dry volume ppm)
O2
20
CO, HC, 100 NO x (dry volume ppm)
O2, CO 2 (dry volume %)
O2
CO, HC, 100 NO x (dry volume ppm)
7.8
22
0
0.50
NOx
16
HC
14
8000
7000
6000
12
5000
10
4000
8
3000
6
4
2000
2
1000
0
0.00 0.05
0.10 0.15 0.20
Equivalence ratio
0.25 0.30
0.35 0.40
Equivalence ratio
Fig. 9. Flue-gas data for the original model configuration shown in Fig. 3a
7
9000
0
0.45 0.50
CO, HC, 100 NO x (dry volume ppm)
3.9
Total thermal input (kW)
In order to try to improve the combustion performance of the combustor, the configuration of the air inlets has been
modified, as shown in Fig. 3b). The measured flue-gas data for this new configuration is represented in Fig. 10. As can
be seen, the effect of the air inlet configuration on both combustor performance and pollutant emissions is marginal.
Total thermal input (kW)
12.0
7.8
O2
20
CO2
18
CO
HC
20
CO2
18
CO
8000
6000
12
5000
10
4000
8
3000
6
2000
4
1000
2
0
0.00 0.05
0.10 0.15 0.20
0.25 0.30 0.35
0.40
0.45
NOx
20.9
HC
14
5000
4000
8
3000
6
2000
1000
2
0
0.00 0.05
0.10 0.15 0.20
0.25 0.30
0.35 0.40
7.8
20.9
25.6
30.1
10000
22
9000
20
CO2
18
CO
18
CO
16
HC
14
8000
7000
O2
6000
12
5000
10
4000
8
3000
6
2000
4
1000
2
0.40
0.45
NOx
O2, CO 2 (dry volume %)
CO2
CO, HC, 100 NO x (dry volume ppm)
20
NOx
O2, CO 2 (dry volume %)
0
0.45 0.50
Total thermal input (kW)
25.6
0.25 0.30 0.35
7000
10
O2
0.10 0.15 0.20
8000
Equivalence ratio
22
0
0.00 0.05
9000
6000
4
0
0.50
10000
12
Total thermal input (kW)
16.3
O2
16
Equivalence ratio
12.0
20.9
9000
7000
14
16.3
22
16
HC
14
7000
5000
10
4000
8
3000
6
0
0.00 0.05
2000
1000
0.10 0.15 0.20
0.25 0.30
0.35 0.40
Equivalence ratio
Fig. 10. Flue-gas data for the modified model configuration shown in Fig. 3b
8
8000
6000
2
Equivalence ratio
9000
12
4
0
0.50
10000
0
0.45 0.50
CO, HC, 100 NO x (dry volume ppm)
O2, CO 2 (dry volume %)
NOx
16
12.0
10000
O2, CO 2 (dry volume %)
22
CO, HC, 100 NO x (dry volume ppm)
7.8
CO, HC, 100 NO x (dry volume ppm)
Total thermal input (kW)
5.8
4. CONCLUDING REMARKS
The flow characteristics of a combustor model under non-reacting conditions have been investigated using LaserDoppler Anemometry. Data is reported for mean and turbulent velocities as a function of the air mass flow rate and its
preheating temperature. The main conclusions are as follows: i) a common feature to all test conditions is the
establishment of a large recirculation zone; ii) mean and turbulent velocities increase within the recirculation zone as
the air mass flow rate increases; iii) near the combustor outlet the mean and turbulent velocities are higher for the
intermediate air mass flow rate tested and iv) the effect of the air inlet preheating temperature on the flow field is
marginal. The isothermal flow characterization was followed by combustion measurements at the exit of the combustor
model. Measurements of mean gas species concentration (O2, CO2, CO, HC and NOx) are reported as a function of the
equivalence ratio and thermal input for two different configurations of the air inlets. The main conclusions are as
follows: i) combustion performance is higher for both lower values of thermal input and equivalence ratio; ii) NOx
emissions are very low regardless of the combustor operating conditions and iii) the effect of the air inlet configuration
on both combustor performance and pollutant emissions is marginal.
ACKNOWLEDGMENTS
Financial support for this work was provided by the European Commission under the contract number ENK5-CT 20000014 and is acknowledged with gratitude. The first author (M. Melo) is pleased to acknowledge the Fundação para a
Ciência e Tecnologia for the provision of a scholarship (SFRH/BD/6345/2001).
REFERENCES
Levy, Y. and Arfi, P. (2004). “Turbulence-Chemistry Interactions Calculations for Improved NOx Predictions”,
Cleanair Journal, in press.
Levy, Y., Sherbaum V. and Arfi, P. (2004a). “Basic Thermodynamics of FLOXCOM, the Low-NOx Gas Turbines
Adiabatic Combustor”, Applied Thermal Engineering, in press.
Levy, Y., Banzger (Polyakh), Y. and Sherbaum V. (2004b). “Theoretical Investigation of Single-Point Water Injection
In Cross-Flow”, submitted for publication.
Melo, M., Sousa, J. and Costa, M. (2002). “PDA Measurements of Single Point Injection in Cross-flow”, 11th
International Symposium on the Application of Laser Techniques to Fluid Mechanics, Lisbon, 7-11 July.
9