- Email: [email protected]
CO synthetic gas for coal integrated gasification combined cycle applications
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications Min Chul Lee a,1, Jisu Yoon b, Seongpil Joo b, Youngbin Yoon b,* a
Department of Safety Engineering, College of Engineering, Incheon National University, Academy-ro 119, Yeonsugu, Incheon 406-772, Republic of Korea b Department of Mechanical and Aeronautic Engineering, Seoul National University, Gwanack-Gu, Seoul 151-742, Republic of Korea
article info
abstract
Article history:
This paper describes the gas turbine combustion characteristics of coal-derived synthetic
Received 16 December 2014
gas (syngas), particularly for the syngases at the Taean IGCC plant in Korea and the Bug-
Received in revised form
genum IGCC plant in the Netherlands. To evaluate the combustion performance of these
13 June 2015
syngases, we conducted combustion tests with elevated temperature and ambient pres-
Accepted 16 June 2015
sure in a GE7EA model combustor. We observed flame stability, dynamic pressure char-
Available online xxx
acteristics, NOx and CO emissions, temperature in the combustion chamber, and flame structures while varying the heat input and diluent integration ratio. Without dilution,
Keywords:
Buggenum's stable regime is larger than that of Taean, since a higher hydrogen content
Integrated gasification combined
causes sustained flames, even at a very high flame stretch rate. However, when consid-
cycle
ering nitrogen dilution, Taean's syngas has a more stable burn than that of Buggenum,
Combustion performance test
since an increase in nitrogen at Buggenum has negatively affected flame stability. From the
Synthetic gas combustion
results of NOx/CO emissions and combustion efficiency, we report that both syngases are
Flame stability
sufficient to control NOx emissions at under 5 ppm with almost complete and stable
Combustion instability
combustion; however, quantitatively, the effect of nitrogen dilution is slightly different at
NOx emission
Taean and Buggenum due to slight differences in the H2/CO ratio and diluent heat capacity. All the tested results and conclusions drawn are considered for optimal operation and trouble shooting at the Taean IGCC plant, which is scheduled to be complete toward the end of 2016. Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ82 2 880 7396; fax: þ82 2 872 8032. E-mail addresses: [email protected] (M.C. Lee), [email protected] (Y. Yoon). 1 Tel.: þ82 32 835 8295; fax: þ82 32 835 0779. http://dx.doi.org/10.1016/j.ijhydene.2015.06.086 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Abbreviations C cp,i CCS Dh Dswirl_in Dswirl_out DRN2 hc IRN2 FLmix FLi IGCC LHV LFL m_ i m_ air m P3 P0ð3Þ; rms PIV ppm f Rair Re r Sn SCGP T4 T1 Tad t T3 q Ujet UHC Vpz xi [X]
Sonic velocity in the air at the temperature of T Specific heat at constant pressure of i-gas Carbon capture and storage Hydraulic diameter Inner diameter of swirler Outer diameter of swirler Dilution ratio of nitrogen Combustion efficiency Integration ratio of nitrogen Flammable limit of mixture Flammability limits of binary sub-mixtures Integrated gasification combined cycle Lower heating value Lower flammable limit Mass flow rate of i-gas Mass flow rate of air Dynamic viscosity Combustion chamber inlet pressure Root mean square of the dynamic pressure at the dump plane Particle image velocimetry Parts per million Equivalence ratio Ideal gas constant for air Reynolds number Density Swirl number Shell coal gasification process Combustor liner temperature Air inlet temperature Adiabatic flame temperature Residence time Combustion chamber inlet temperature Swirl vane angle Jet velocity Unburned hydro carbon Volume of the primary zone Volume percentage of a sub-mixture in total fuel Molar concentration of X species
Introduction Background IGCC is a potential technology for addressing the growing concern regarding global warming, urban atmospheric pollution, and depletion of energy resources. IGCC uses abundant low rank coal or renewable fuels such as biomass and agricultural or municipal waste, and offers environmentally clean benefits such as low NOx combustion, syngas desulfurization, and CCS capabilities. Although this promising technology has been developed and demonstrated since 1990 at numerous sites (e.g., Wabashi River and Tampa in the US, Puertollano in Spain, Buggenum in the Netherlands and Nakoso in Japan),
low plant availability in initial operation years is considered a serious issue. More precisely, unplanned outages due to fuel nozzle and liner burn out, combustion instability, and compressor surging in gas turbine units are the most critical sources of low availability [1e3]. Since we expect many outages in gas turbine combustors at new IGCC plants, due to the conditions at existing IGCC plants, a more detailed understanding of syngas combustion is necessary, and advanced syngas turbine technology should be adapted to improve IGCC production. The first Korean IGCC project (power output: 300 MWe; plant site: Taean) was launched in 2006 and is expected to be complete toward the end of 2015. For better operational availability and combustion optimization at this plant, we tested Taean syngas in a model GE7EA gas turbine combustor in this study to quantify potential application problems and evaluate its combustion performance as a gas turbine fuel.
Prior research on syngas turbine combustion Due to the growing interest in IGCC during the last two decades, many researchers have studied syngas combustion. These syngas combustion studies are divisible into two categories: fundamental research on combustion characteristicsdsuch as laminar burning velocity [4,5], ignition [6,7], flashback [8], blowout [9], flammable limit [10], and emissions [11e13]dand advanced research using gas turbine-like combustors under gas turbine-similar conditions [14e20]. As researchers in the first category, Natarajan et al. (2007) examined the effect of CO2 dilution, preheating, and pressure on the laminar flame speed of H2/CO mixtures by using two measurement approaches: flame area images of a conical Bunsen flame and velocity profile measurements of a onedimensional stagnation flame [4]. Walton et al. provided an experimental data set on the ignition delay time of hydrogen and carbon monoxide syngases under high pressure and high temperature conditions [6]. From the regression analysis of their data set, they obtained the best-fit correlation between ignition delay time and experimental variables. Furthermore, Lieuwen et al. investigated the impact of syngas composition on four combustor-operability issuesdblowout, flashback, combustion instability, and autoignitiondby analyzing calculations obtained from the CHEMKIN program [9]. Giles et al. investigated the NOx emission characteristics of syngas mixtures with airstream dilution of H2O, CO2, and N2 in a counterflow diffusion flame. They found that the NOx reduction effect of dilution was H2O > CO2 > N2 [12]. Such scientific foundations have provided useful insights and information to use when analyzing results from advanced studies, which enables the development of rather economical approaches to understand combustion phenomena in syngas combustors before conducting complicated tests on full-scale engines. Researchers in the second category have conducted studies, occasionally at great expense, to evaluate the combustion performance of syngas combustors and investigate syngas combustion characteristics in simulated gas turbine environments. Littlejohn et al. conducted combustion tests on a prototype injector of Taurus 70 engine for various syngas compositions and determined the lean blowout limits, loglinear dependency of NOx, and the turbulent flame speed
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
3
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
correlation using PIV data [14]. Dodo et al. investigated the performance of a multiple-injection, dry low NOx combustor with hydrogen-rich syngas fuels and reported the effectiveness of the convex perforated plate swirler on enhanced cooling of the combustor liner and suppression of combustion oscillation and NOx [15]. Through combustion tests of a can combustor of heavy-duty gas turbines, Hasegawa et al. reported many useful results related to NOx/CO emissions, pattern factors, combustion efficiency, and thermal efficiency [16e18]. Lee et al. also tested a can combustor for a GE7EA gas turbine combustor and reported the combustion performance characteristics of synthetic natural gas and H2/CO syngas with N2, CO2, and steam dilution [19e21]. They found that a 40% dilution of N2, CO2, and steam reduced NOx emission by 79%, 88%, and 95%, respectively, and that the NOx reduction per unit power is logarithmically related to the diluent's heat capacity. Although the abovementioned practical studies achieved technical progress in syngas turbine combustion, many unsolved problems remain due to the highly nonlinear behavior of combustion characteristics according to syngas composition, combustor geometry, and test conditions. In addition, our knowledge of syngas combustion in commercial gas turbines is mostly confined to commercial gas turbine makers, such as GE and Siemens, and operators of existing plants. Motivated by this consideration, we tested the Taean syngas to obtain a broad understanding of combustion characteristics under conditions relevant to a gas turbine. In particular, in this paper, we discuss the results from the combustion performance test in terms of flame stability, thermo-acoustic combustion instability, NOx and CO emissions, combustion efficiency, and flame structure. Furthermore, we compare this with the Buggenum syngas, which uses the same gasification technology as TaeandSCGP. The ultimate goal of this study is to provide appropriate solutions to syngas combustion issues and improve the efficiency and reliability of the Taean IGCC plant through an understanding of syngas combustion characteristics.
Test method The gas turbine combustion test facility and operating conditions For the purposes of this study, an atmospheric pressure combustion test facility for 60 kW-scale gas turbines was installed, as shown in Fig. 1. This facility consisted of an air compressor, air heater, a cooling and combustion air feed line, an atmospheric pressure combustor, an external stack with a silencer, and a fuel supplying system. Coal derived syngases are mainly composed of H2, CO, N2, CO2 and steam, and last three gases are incombustible diluents of which heating value is zero. Since these diluents are almost interchangeable in views of NOx/CO emission and combustion instability when the diluent's heat capacity is same [21]. Thus, only N2, the largest content in syngas, has been selected as a representative diluent to simplify the test. The fuel supply system can control the flow rate of H2, CO, and N2, respectively and both the fuel and air were processed to meet the choking condition immediately upstream of the combustor to prevent the perturbation of air and fuel in the combustion chamber by blocking the influence of pressure variation on the upstream flow. As shown in Table 1, the tests were conducted at a slightly elevated pressure, since a water-cooled plug nozzle blocked Table 1 e Experimental conditions. Items
Unit
Combustion chamber pressure Air inlet temperature Mean mixture jet velocity at combustor inlet Reynolds number Swirl number Heat input Overall equivalence ratio Nitrogen dilution ratio
Air compressor Regulator
Value
Bara C m/s
1.1e1.4 200 41.2e74.9
e e kWth L/L %
28,398e51,600 0.832 30e60 0.381e0.766 0e150
Flow FCV meter
Dryer
MFC
Filter
H2
Air heater MFC
Cooling air
CO
Combustion air
Regulator
Silencer Stack
N2 for purge Regulator
N2
MFC
GE7EA combustor
Gas sampling
Inline static mixer Fig. 1 e Schematic view of gas turbine combustion test facility.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
To offer a simpler description, the DRN2 and IRN2 are defined as shown in the following equations:
90% of the combustion outlet area to form an acoustic boundary for the combustion chamber outlet. The inlet temperature of combustion air was controlled at 200 C to meet conditions relevant for a gas turbine. According to the fuel and airflow rate, the average jet velocity of the fuel-air mixture at the combustor inlet varied from 41.2 m/s to 74.9 m/s, which are values that are comparable with those of a real engine. The Reynolds number, calculated by the following equation, was maintained between 28,398 and 51,600: Re ¼
rUjet Dh ; m
DRN2 ¼
IRN2 ¼
flow rate of N2 ½slpm : N2 equivalent dilution ratio total fuel flow rate ½slpm
where the N2 equivalent dilution ratio can be determined from Equations (2 and 3) by obtaining the m_ N2;eq of each syngases and substituting it into the flow rate of N2. DRN2 indicates the amount of nitrogen contained in the fuel compared to the combustible H2/CO gas, and IRN2 indicates the amount of nitrogen content in the fuel compared to the composition of the Taean or Buggenum syngases. That is, 100% of IRN2 is 56.4 mol % and 65.4 mol% nitrogen in the fuel at Taean and Buggenum, respectively. This represents the conditions under which all nitrogen produced from an air separation unit is supplied to the gas turbine combustor and, thus, when an IGCC plant is operating with the maximum integration of nitrogen.
(1)
m_ CO2 cp;CO2 þ m_ steam cp;steam þ m_ Ar cp;Ar ; cp;N2
(3)
(4)
where r, Ujet , Dh, and m are density of mixture, jet velocity of mixture, hydraulic diameter at combustor inlet, and dynamic viscosity, respectively. The turbulent Re was 45,000e81,000, assuming a relative turbulence intensity of 10% and an integral scale of 1/10 in the chamber diameter. Although chemical reaction times vary significantly according to the fuel composition and the amount of nitrogen dilution, the operational points in this research lie between the wrinkled laminar-flame regime and the flame in-eddies regime in the € hler-Turbulence Reynolds chart, as is true for other gas Damko turbines and internal combustion engines [22]. Fuel was supplied from each bottle of feedstock gases while keeping fuel composition constant and varying heating values from 30 kW to 60 kW, the span of which is 5 kW. Fuel gases were mixed well through the inline static mixer and supplied to the combustor. As previously described in the explanation on the fuel supply system, only nitrogen, instead of carbon dioxide and steam, was used as a diluent because the effect of dilution on combustion performance is almost the same when diluents have the same heat capacity [21]. Thus, the equivalent mass flow rate of nitrogen is determined from the following equation: m_ N2 ;eq ¼ m_ N2 þ
flow rate of N2 ½slpm : sum of flow rate of combustible gases ½slpm
Model gas turbine combustor To examine the combustion characteristics of the syngases, a 1/3 scaled-down one-can dump combustor of a GE7EA gas turbine was designed and fabricated as a test bed combustor. Fig. 2 shows the schematic of the combustion test rig. The combustion chamber consisted of two parts. The first part was made from optically accessible quartz, which was cooled by a high-pressure injection of the same amount air as the combustion air, and the second part was made from steel. To meet gas turbine relevant conditions, combustion air at 356 ± 5 C controlled by an mass flow controller (MFC, Bronkhorst F206BI, uncertainty ¼ ±0.8%) was supplied to the flame through a central annular swirling nozzle (swirl number ¼ 0.832, 14 swirl channels, i.d. ¼ 25 mm, o.d. ¼ 40 mm). Tests were conducted at a slightly aviated pressure (1.1e1.4 bar), since a 90% area of the combustor outlet was blocked by a water-cooled plug nozzle to form an acoustic boundary at the combustor
(2)
where m_ i and cp,i are mass flow rate and specific heat at constant pressure of i-gas.
Static pressure sensor Dynamic pressure sensor Temperature sensor T(4)
T(1) P
(1)
P
Burner head (2)
Fuel nozzle
P
(4)
P SP(2)
(5)
P
(6)
P
T(5) (7)
P
(8)
P SP(2)
(9)
P
(10)
P
(11)
T(2)
P
(3)
SP(1)
Ignitor T(3)
180 mm (Combustor liner - quartz part)
1230 mm (Combustor liner - steel part)
Plug nozzle
Fig. 2 e Schematic diagram of the gas turbine combustor and test rig with measurements. Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
5
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
outlet. The detailed description on combustor used for this study is given in Ref. [23].
Table 2 e Syngas composition at Taean and Buggenum [3].
Measurements
Syngas composition
As shown in Fig. 2, eleven piezoelectric dynamic pressure sensors (PCB 102A05), three static pressure sensors, and five ktype thermocouples were installed at the combustor and air inlet line to measure the dynamic and static pressure and temperature of each location. To minimize the measurement error of temperature at combustor due to radiative heat transfer from thermocouple to combustor wall, temperature has been measure at the thermally stabilized condition. However, thermocouples are not radiation-shielded so an error around 5% of measured temperature can be introduced. Exhaust gas was sampled at the combustor outlet, and the concentrations of NOx, CO, O2, CO2, and UHC were measured by a TESTO 360 gas analyzer. To measure the NOx and CO emissions, the exhaust gas was sampled at the exhaust end of the combustor and analyzed using a TESTO 360 gas analyzer. The measured results for emissions were corrected to 15% oxygen on a dry basis using the following equations [24]
H2 CO CH4 N2 CO2 H2O Ar H2S Sum Higher heating value Lower heating value H2/CO Mole fraction of combustible gas Lower flammable limit in air Upper flammable limit in air N2 equivalent dilution ratio Upper Wobbe index Lower Wobbe index
ðmeasured NOx Þ ðNOx at 15% O2 Þ ¼ 29:79 ) ( 4:76 ð2 ðmeasured O2 ÞÞ 1 1 4:76 ðmeasured O2 Þ ðmeasured CO Þ ðCO at 15% O2 Þ ¼ 29:79 ) ( 4:76 ð2 ðmeasured O2 ÞÞ 1 1 4:76 ðmeasured O2 Þ
Unit Taean Buggenum Difference [A] [B] [(A B)/A] mol% 12.9 mol% 31.5 mol% 0.1 mol% 50.8 mol% 0.4 mol% 4.2 mol% 0.4 mol% 0.0 mol% 100.0 MJ/Nm3 5.60 5.35 MJ/Nm3 % 40.6 mol% 44.3
12.3 24.8 0.0 42.4 0.8 19.1 0.6 0.0 100.0 4.70 4.46 49.6 37.1
3.54% 21.12% 0% 17.09% 114.57% 341.86% 37.09% 0% 0% 16.07% 16.64% 22.47% 16.06%
%
19.1
21.2
10.99%
%
73.3
71.7
2.18%
%
56.4
65.4
15.96%
MJ/Nm3 MJ/Nm3
10.02 9.57
9.44 8.95
5.79% 6.48%
(5)
(6)
To investigate flame structure, an OH*-chemiluminescence image of flame was obtained using a PI-MAX Gen II ICCD camera with two optical filters (UG11, WG305) and a UVNikkor 105 mm lens. Because OH*-chemiluminescence images are line-of-sight integrated images, a three-point Abel deconvolution scheme was used to extract two-dimensional information from the line-of-sight integrated images.
Results and discussion Combustion tests on both fuel gases were conducted with respect to the heat input and DRN2, and the results were compared to understand the effect of slight differences in syngas composition on combustion performance. We also compared the results to predict various combustion phenomena that occur when Taean or Buggenum syngas fuels a GE7EA gas turbine. Table 2 outlines the fuel composition of Taean and Buggenum, which were determined from the reports of Taean and Buggenum IGCC plants [3]. The main features of the Taean syngas compared with the Buggenum syngas are a lower H2/CO ratio, lower diluents in fuel due to low moisture, and thus higher heat content per unit volume.
Flame stability map In the present study, we classified the stability regime according to the flame shape and intensity of pressure
fluctuation, as reported by many flame stability map researchers [25e27]. Fig. 3 describes the criteria for classification on the flame stability map. Regime I is a V-shaped flame; this flame shape occurs due to swirling flow and outwardly diverging recirculation flow toward the dump plane when the flame root is attached to the fuel nozzle. Regime II is an Mshaped flame; as opposed to regime I, this type of flame interacts with a vortex flow and burns close to the dump plane and quartz liner. In regime IV, a long cone-shaped flame occurs, and it seems like a long tornado due to the swirling flow. In regime III, flames of regimes II and IV occur alternately, and this flame oscillation results in large pressure fluctuations. In regimes V and VI, lift off and blowout phenomena occur, respectively. Fig. 4(a) shows the flame stability map of Taean and Buggenum syngas according to DRN2. For both types of fuel, flame changes from strong burning regime I through II, III, IV, and V to regime VI as the DRN2 increased and the heat input decreased. Regime I occurred at just over 40 kW of heat input with a DRN2 of under 100%. As the nitrogen dilution increased, more enhanced recirculation flow occurred due to the high fuel-air mixture momentum, which resulted in the flame's transition from regime I to regime II. Further nitrogen additions in syngas slowed down the combustion reaction, and consequently lowered the burning velocity. Moreover, since a nitrogen addition increases the mean mixture velocity, the flame was extinguished locally, being detached from the dump plane and moving downstream when the mean mixture velocity exceeded the turbulent burning velocity. This movement downstream resulted in a regime shift from II to VI, passing through a narrow region of III, in which the flame oscillated rather unstably with an alteration motion between detachment and attachment to the dump plane. Next, flame lifting (regime V) and lean blowout (VI) phenomena occurred
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Regime
Flame images
Description
V-shaped flame I
(Flame does not interact with outer recirculation flow)
M-shaped flame II
(Flame interacts with outer recirculation flow)
Oscillating flame (Flame oscillates alternately between M-shaped flame and
III
long cone-shaped flame)
Long cone-shaped flame (Flame does not interact with outer recirculation flow and
VI
seems like a tornado due to swirling flow)
Lift-off
V
VI
Blowout
No Flame
Fig. 3 e Criteria for classification on a flame stability map.
sequentially with decreasing heat input and increasing nitrogen dilution. Combustion instability did not occur for all tested conditions without regime III (detailed results and discussions on the combustion instability characteristics are in section 5 of chapter 3). Therefore, when considering only the flame stability map, we conclude that only regimes I and II are stable operating conditions for Taean and Buggenum syngas combustion, and the Taean syngas is more stable when compared to the Buggenum syngas since operating condition of Taean is located farther from the unstable regime III than that of Buggenum. In the meantime, we redrew the flame stability map according to the nitrogen mole fraction in fuel, as depicted in Fig. 4(b). This figure illustrates several important points missing from Fig. 4(a). First, in the low load condition, excessive nitrogen dilution (exceeding 30 mol%) can cause unstable
combustion. Second, all borderlines (except the border between regimes I and II) asymptotically approach the range between 77 mol% and 82 mol%. Third, the flames of both fuels cannot exist when the nitrogen content is over 82 mol% in the fuel. To understand why a flame blows out over a certain limit, we estimated the LFL of the H2/CO/N2 mixture on the basis of the Le Chatelier rule [28,29]: FLmix ¼
x1 FL1
100 ; x3 x2 þ FL þ FL þ / 2 3
(7)
where FLmix is measures the flammability limit of the mixture in vol%, FLi is the flammability limit of the binary submixtures comprised of one combustible gas and one inflammable gas, and xi is the volume percentage of a sub-mixture in P total fuel ( xi ¼1). Assuming that the sub-mixtures are H2 þ N2 and CO þ N2, we calculated the maximum LFL of
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
7
Fig. 5 e Temperature characteristics of Taean and Buggenum syngases with respect to the nitrogen integration ratio.
Fig. 4 e Flame stability map of Taean and Buggenum syngases.
Taean and Buggenum syngases to be 83.9 vol% and 84.5 vol%, respectively. These value lines are plotted in Fig. 4(b), and the blowout limit occurred because both fuels approach the LFL lines of syngas but cannot exceed them. It must be noted that even though a high swirl stabilizes a flame almost until the LFL, the flame cannot go beyond the maximum LFL of syngas. Inversely, the blowout limit line of any mixture can be estimated by obtaining the maximum LFL of a mixture.
Temperature characteristics Fig. 5 presents measurements of T1 and T4 with respect to IRN2. With these temperatures, the Tad, calculated via CANTERA code with the GRI 3.0 mechanism, is plotted. For all test
conditions, the T1 is maintained at 200 ± 3 C, which indicates that the air preheater controls T1 well and that the test's accuracy in terms of temperature is over 98.5%. As expected, T4 and Tad show a linearly increasing tendency as heat input increases because the flame temperature increases following the rise in heat input. As the IRN2 increases, the Tad decreased almost linearly since the addition of inflammable nitrogen increased the total heat capacity in the fuel and, thus, the heat demand to raise the flame temperature. The nitrogen-induced flame cooling effect at Buggenum is stronger than that at Taean, since the Buggenum syngas has a higher nitrogen mole fraction in the fuel, as shown in Table 2. This fact implies that the Buggenum IGCC plant expends 16% more nitrogen to control NOx emissions, thereby resulting in a 8.2% larger decrease in Tad at 60 kW with 100% DR (Tad of Taean has decreased from 2066 C to 1627 C whereas Tad of Buggenum has decreased from 2071 C to 1507 C). We observed similar characteristics in T4 for the same reasons. However, we observed a sudden drop in T4 at over 50% of IRN2. Flame position and structure largely affected T4, so the stability mode change from regime II to regime IV via regime III resulted in the change in T4. For example, for the Taean syngas at 30 kW, we observed a sudden drop between 39% and 50% of IRN2, and this condition corresponds to the 47.3% and 86.4% of DRN2 in regimes II and IV, respectively, according to Fig. 4(a). On the other hand, for the Buggenum syngas at 30 kW, we observed a sudden temperature drop between 50% and 100% of IRN2, and this condition corresponds to the 88.2% and 176.3% of DRN2 in regimes III and IV, respectively. Thus, it can be concluded that flame movement downstream is the main cause of a sudden drop in combustor liner temperature. From this result, it can be concluded that nitrogen for NOx reduction should be carefully controlled in a lower load (under 40 kW) since flames are easily destabilized. Consequently, the flame might lift off and then damage downstream portions of the combustor, such as turbine blades and vanes.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
8
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
NOx and CO emission characteristics Fig. 6(a and b) depict the emission characteristics of NOx and CO with respect to IRN2. For both syngases, the concentration of NOx emissions showed an exponentially decreasing trend when adding more nitrogen to the fuel for all heat inputs. We attributed these results to the decrease in flame temperature via increased nitrogen integration and the extended Zeldovich NOx mechanism, which explains that NOx emissions exponentially correlate with flame temperature [30]. From the three Zeldovich reactions, assuming a quasi-steady state of nitrogen atoms, we obtained the following NO-formation equation d½NO ¼ 2k1 ½O½N2 ; dt
(8)
where the rate coefficient k1 ¼ 1.8∙1014 exp (318 kJ mol1/ (RT)) [cm3/mol-s], and [X] is the molar concentration of X species. Thus, NO can be minimized by decreasing [N2], [O], or k1 (i.e., by decreasing the flame temperature). It is noteworthy that increased N2 enhances NOx emissions in view of [N2] and simultaneously suppresses NOx in view of flame temperature. However, flame temperature dominates the NOx emission, and thus NOx decreases as N2 increases because k1 is insignificant at temperatures below 1700 K [31]. This feature is evident in Fig. 7, which illustrates the NOx emission dependence on Tad which has been already mentioned in chapter 3.2. For both syngases, NOx emissions become significant at Tad > 1450 C and increase exponentially as Tad increases. Furthermore, most of the test data fit well in the exponential asymptotic curve. This good fit implies that Tad mainly affects NOx, not the nitrogen mole fraction. Discordance and early rise of NOx in low heat input conditions might be attributed to the nature of a partially premixed flame, which forms a localized high temperature zone. NOx emissions of the Taean syngas were slightly higher than that of Buggenum by 2.09 ppm at 60 kW without N2 dilution. Since flame temperature is almost the same for that condition, residence time in flame can be considered as the main cause of the NOx emission difference. We calculated the residence time using the following equation, which is based on hot gas flow through the primary flame zone [32,33]: t¼
Vpz P3 ; m_ air Rair T3
Fig. 6 e NOx and CO emissions of Taean and Buggenum syngases.
(9)
where Vpz, P3, m_ air , Rair, and T3 are volume of the primary zone [m3], combustion chamber inlet pressure [Pa], mass flow rate of air [kg/s], ideal gas constant for air [J/kg-K], and combustion chamber inlet temperature [K], respectively. Herein, P3, m_ air , Rair, and T3 are the same for both syngases and only Vpz is a control variable of t, which is obtained from the timeaveraged Abel-deconvoluted OH*-chemiluminescence images in Fig. 8. The calculated Vpz of the Taean syngas was 5020 mm3, which was larger than that of Buggenum by 8.2%; consequently, the t of the Taean syngas was longer than that of Buggenum by 4%. Therefore, we can assert that an 8.2% longer t in a flame is a vital cause of increased NOx emissions, when comparing the Taean and Buggenum syngases. This volume difference might be due to the H2/CO ratio in the fuel. Since H2 burns more than 10 times faster than CO (see
Fig. 7 e NOx emissions of Taean and Buggenum syngases with respect to the adiabatic flame temperature.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
9
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Fig. 8 e Primary combustion zone of Taean and Buggenum syngases (left-half images of time-averaged Abel-deconvoluted OH*-chemiluminescence).
Table 3), the Taean syngas's flame propagation speed, which contains more CO, must be lower than that of the latter. Hence, this relatively slow reaction resulted in a larger Vpz and longer t. For both tested syngases, the concentration of CO was very high, reaching up to 500 ppm or more for lower loads with high nitrogen integration (over 50%), but the concentration decreased sharply for higher loads (down to 10 ppm or less), showing almost perfect combustion (combustion efficiency > 99.9%). Because the high emissions of CO stemmed from the low mixed mean adiabatic combustion temperature, the nitrogen injection (which quenches flame) can be a main source of high CO emissions for the low load. When increasing the load, sufficient main fuel supply can cause the flame temperature to rise, and thus CO emissions drop to a sufficiently lower level. Fig. 9 illustrates the relationship between CO emissions and flame temperature. CO is rapidly emitted
when Tad < 1230 C, and this spiking relationship between CO and flame temperature coincides with the results from Yoshimura et al. [31], who reported that CO emissions exponentially increase at Tad < 1500 K. By comparing the Buggenum and Taean syngases, we found that the Buggenum syngas is prone to emit more CO. For example, over 1000 ppm of CO is emitted when firing the Buggenum syngas at 40 kW with 150% IRN2. On the other hand, we detected no significant CO emission under the same conditions for the Taean syngas. This discrepancy can be explained by the fact that the Buggenum IGCC plant integrates more nitrogen to control NOx, as shown in Table 2, because the Tad at Buggenum is significantly lowered when IRN2 increases and the slope is steeper than that of the Taean syngas, as shown in Fig. 6(a). From the NOx and CO emission results, we conclude that emissions characteristics highly correlate with flame temperature, which relates to IRN2 as well as the H2/CO ratio.
Table 3 e Properties of hydrogen, carbon monoxide, and methane [19,20,33e35]. Property Chemical formula Boiling point Specific gravity Flammable limits in air Autoignition temperature Max. burning velocity Max. adiabatic flame temperature Stoichiometric air/fuel ratio Higher heating value Lower heating value Upper Wobbe index Lower Wobbe index
Unit
Hydrogen
Carbon monoxide
Methane
e C (vs. air) % C cm/s C kg/kg MJ/Nm3 MJ/Nm3 MJ/Nm3 MJ/Nm3
H2 252.8 0.07 4e75 500 289 2200 34.1 12.76 10.80 48.23 40.65
CO 191.5 0.97 12.5e74.2 609 19 2210 2.5 12.63 12.63 12.80 12.80
CH4 161.5 0.55 5.0e15.4 537 37 2000 16.9 39.75 35.82 53.28 47.91
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
10
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
lower flame temperature, and these emissions were suppressed to under 10 ppm when the heat input was over 45 kW.
Combustion oscillation characteristics
Fig. 9 e CO emissions of Taean and Buggenum syngases with respect to the adiabatic flame temperature.
However, even though an increased nitrogen molar concentration should enhance NOx formation, according to the extended Zeldovich mechanism, this effect was not significant in our study when compared to the effect of flame temperature. When considering only the H2/CO ratio, the Taean syngas emitted more NOx than the Buggenum syngas due to a larger t; these emissions were suppressed to under 3 ppm when IRN2 was over 50%. On the other hand, when increasing the IRN2, the Buggenum syngas emitted more CO because of a
Fig. 10 illustrates our investigation of combustion oscillations of both syngases with respect to IRN2. The combustion pulsation level was very low in all tests (i.e., there was almost zero combustion instability without dilution in the case of 35 kW with 100% IRN2). This low fluctuating level is much lower than the oscillation in a standard gas turbine combustor, and it is at least 5% under the control limit for oscillation in a standard turbine combustor. These results are comparable to the results of previous tests on H2 and CO combustion with and without dilution [20,21]. To understand why combustion instability did not occur, two mechanisms of particular significance in gas turbine combustion systems can be introduced: oscillation feedback among acoustic pressures, equivalence, and heat release [34,35] and flame-vortex interaction [36,37]. In the former mechanism, an acoustic pressure fluctuation in the combustor results in a velocity and acoustic pressure fluctuation at the fuel nozzle, which perturbs the fuel flow rate. This perturbation propagates to the flame, where it produces a heat release oscillation and, in turn, another pressure fluctuation. Here, heat release oscillation is the periodic repetition of extinction and re-ignition, and a syngas with considerable hydrogen content cannot be easily extinguished due to the large flammable limit and flame extinction
Fig. 10 e RMS values of pressure fluctuations with nitrogen integration ratio (Upper: Taean syngas; Lower: Buggenum syngas). Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
stretch rate of highly reactive hydrogen [38]. Thus, fluctuations in heat release are minimal and do not cause fluctuations in pressure or fuel flow rate. In the latter mechanism, large-scale, coherent vortical structures cause an oscillation in the rate of heat release. These structures are the result of flow separation from flame holders and rapid expansions, as well as vortex breakdown in swirling flows. However, in this study, neither a vortical flame structure at the outer recirculation zone nor flow separation near the fuel nozzle is observed (as shown in Fig. 12) since hydrogen (whose burning velocity is very high) causes the flame shape to be diffusive and short. In the case of 35 kW with 100% IRN2, we observed a large amplitude pressure fluctuation in both syngases. Thus, to analyze the oscillation frequency, we plotted Fig. 11 by conducting the Fast Fourier Transformation. The main frequency was 10 Hz, which was due to flame oscillation in regime III, and 300 Hz and higher frequencies represented fundamental longitudinal frequency and their harmonics. Therefore, only flame oscillation near flame lift-off conditions can be considered as a significant source of combustion instability for both syngases, and flames in regimes I and II were not a concern for thermo-acoustic instabilities.
Flame structures Fig. 12 illustrates the time-averaged OH*-chemiluminescence images (upper half) and their Abel-deconvoluted images (lower half). Since the flow is assumed to be axisymmetric, only half images are shown. In low heat input conditions, the flame is weak and broader, because the outer recirculating flow can easily affect a flame at a low equivalent ratio. When increasing the heat input, a flame grows larger and stretches downstream, and the primary combustion zone moves to the center due to increased momentum in the downstream direction. When increasing IRN2, the burning intensity in both syngases diminishes and the flame moves from upstream/center to downstream/outer wall of the combustion chamber, thereby increasing the combustion region. The high temperature combustion region in which the thermal NOx occurred
Fig. 11 e FFT of pressure fluctuations of Buggenum syngas at 35 kW with 100% nitrogen dilution.
11
contracted substantially, but the lower temperature combustion zone expanded. As previously described in subchapter 3.3, these flame structure results coincide with the results of NOx and CO emissions.
Conclusion We conducted combustion performance tests of the Taean and Buggenum syngases in a 1/3 scaled-down GE7EA industrial gas turbine combustor. We obtain the following conclusions from the test results: (1) We plotted the flame stability map by categorizing flames into six regimes according to the flame shape and pressure fluctuation. We observed stable combustion in regimes I and II, into which both Taean and Buggenum syngases fall. When considering only the H2/ CO ratio in the fuel, the Buggenum syngas is more stable. However, when considering all aspects, including IRN2, the Taean syngas is more stable. Furthermore, flames from both syngases cannot exist when the nitrogen level is over 82 mol%, and this content almost corresponds to the maximum LFL estimated by the Le Chatelier rule. Therefore, we concluded that the lean blowout limit of syngas is predictable by simply calculating maximum LFL, even in the high swirl stabilized flame. (2) Temperature characteristics with respect to IRN2 and heat input showed monotonic increasing or decreasing trends, except at certain unstable flame conditions. In these unstable conditions, we observed sudden drops and elevations in combustor temperature due to flame movement. This condition should be avoided since it can damage the combustor's liner and transition piece as well as diminish the lifetime of hot gas path components, including turbine blades and vanes. (3) NOx and CO emissions highly correlate with flame temperature, and the optimal operating conditions of syngas turbines is determined to be when 1230 C < Tad < 1500 C. Under the same flame temperature, the Taean syngas, which contains more CO in the fuel, emits more NOx than the Buggenum syngas does due to a larger t at the former's primary combustion zone. On the other hand, the Buggenum syngas, which contains more diluents, emits more CO because of a lower flame temperature. We verified that fuel-side nitrogen dilution is very effective in suppressing NOx emission by lowering the flame temperature, and NOx emission decreases to under 3 ppm when IRN2 is over 50%. However, operating with a low load and a high DRN2 caused problems with CO emission and flame stability. Thus, the dilution system of an IGCC plant should be operated with analysts cautiously monitoring the CO emission and flame stability for better fuel efficiency and plant reliability. (4) Combustion instability occurred only near regime III due to the flame attachment/detachment mechanism. Thus, operating conditions, including IRN2, should avoid this combustion instability. In the other regimes, we did
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
12
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Fig. 12 e OH*-chemiluminescence images (upper) and Abel-deconvoluted images (lower) for Taean and Buggenum syngases.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
not observe combustion instability for two reasons: 1) lower heat release oscillation and its feedback to pressure fluctuation, and 2) flame shape, which does not generate vortical flow. (5) We used flame structure to analyze the test results of combustion performance, such as flame stability, NOx and CO emissions, combustion efficiency, combustion instability, and temperatures in the combustor. All tested data will be used to determine the optimal operating conditions at the Taean IGCC plant and analyze the plant's outage. (6) Since IGCC with CCS technology is the one of the key technology to solve the global warming problem, many recent demonstration projects such FutureGen in US, GreenGen of China, ZeroGen of Australia have been conducted in the world and Korea government is also planning to construct two more 300 MWe IGCC plants by 2020. However, competing with other new thermal power technology such as ultra-supercritical pulverized coal boiler and circulating fluidized bed technology in views of economics and efficiency and recent large fluctuations of oil and gas price can vary the IGCC perspective in thermal power plant market.
Acknowledgments This work was supported by the Development of 300 MWclass Korean IGCC demonstration plant technology of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant, funded by the Korean Ministry of Trade, Industry and Energy (201195101001C). This work was also supported by the Mid-career Researcher Program, via an NRF grant, funded by the MSIP (NRF-2010-0015100), contracted through the Institute of Advanced Aerospace Technology at Seoul National University.
references
[1] Hornick MJ, McDaniel JE. Tampa Electric Polk power station integrated gasification combined cycle project final technical report. U.S. DOE project 2002; DE-FC21e91MC27363. [2] Dowd RA. Wabash River coal gasification repowering project final technical report. U.S. DOE project 2000; DEFC21e92MC29310. [3] Phillips J. Integrated gasification combined cycle design consideration for high availability. EPRI Technical Report. 2007. p. 1012226. [4] Natarajan J, Lieuwen T, Seitzman J. Laminar flame speeds of H2/CO mixtures: effect of CO2 dilution, preheat temperature, and pressure. Comb flame 2007;151:104e19. [5] Hu E, Huang Z, He J, Jin C, Zheng J. Experimental and numerical study on laminar burning characteristics of premixed methane-hydrogen-air flame. Int J Hydrogen Energy 2009;34:4876e88. [6] Walton SM, He X, Zigler BT, Wooldridge MS. An experimental investigation of the ignition properties of hydrogen and carbon monoxide mixtures for syngas turbine applications. Proc Combust Inst 2007;31:3147e54.
13
[7] Kalitan DM, Petersen EL, Mertens JD, Crofton MW. Ignition of lean CO/H2/air mixtures at elevated pressures. Proc of ASME Turbo Expo. 2006. GT2006:90488. [8] Xu G, Tian Y, Song Q, Fang A, Cui Y, Yu B, et al. Flashback limit and mechanism of methane and syngas fuel. Proc of ASME Turbo Expo. 2006. GT2006:90521. [9] Lieuwen T, McDonell V, Petersen E, Santavicca D. Fuel flexibility influences on premixed combustor blowout, flashback, autoignition, and stability. J Eng Gas Turb Power 2008;130:011506. [10] Noble DR, Zhang Q, Shareef A, Tootle J, Meyers A, Lieuwen T. Syngas mixture composition effects upon flashback and blowout. Proc of ASME Turbo Expo. 2006. GT2006e90470. [11] Whitty KJ, Zhang HR, Eddings EG. Emissions from syngas combustion. Comb Sci Tech 2008;180:1117e36. [12] Giles DE, Som S, Aggarwal SK. NOx emission characteristics of counterflow syngas diffusion flames with airstream dilution. Fuel 2006;85:1729e42. [13] Konnov AA, Dyakov IV, Ruyck JD. Nitric oxide formation in premixed flames of H2 þ CO þ CO2 and air. Proc Combust Inst 2002;29:2171e7. [14] Littlejohn D, Cheng RK, Noble DR, Lieuwen T. Laboratory investigations of low-swirl injectors operating with syngases. J Eng Gas Turb Power 2010;132:011502. [15] Dodo S, Asai T, Koizumi H, Takahashi H, Yoshida S, Inoue H. Performance of a multiple-injection dry low NOx combustor with hydrogen-rich syngas fuels. J Eng Gas Turb Power 2013;135:011501. [16] Hasegawa T, Tamaru T. Gas turbine combustion technology reducing both fuel-NOx and thermal-NOx emissions for oxygen-blown IGCC with hot/dry synthetic gas cleanup. J Eng Gas Turb Power 2007;129:358e69. [17] Hasegawa T, Nakata T. A study of combustion characteristics of gasified coal fuel. Trans ASME 2001;123:22e32. [18] Nakata T, Sato M, Ninomiya T, Hasegawa T. A study on low NOx combustion in LBG-fueled 1500 C-class gas turbine. Trans ASME 1996;118:534e40. [19] Park S, Kim U, Lee MC, Kim S, Cha D. The effects and characteristics of hydrogen in SNG on gas turbine combustion using a diffusion type combustor. Int J Hydrogen Energy 2013;38:12847e55. [20] Lee MC, Seo SB, Chung JW, Kim SM, Joo YJ, Ahn DH. Gas turbine combustion performance test of hydrogen and carbon monoxide synthetic gas. Fuel 2010;89:1485e91. [21] Lee MC, Seo SB, Yoon J, Kim M, Yoon Y. Experimental study on the effect of N2, CO2, and steam dilution on the combustion performance of H2 and CO synthetic gas in an industrial gas turbine. Fuel 2012;102:431e8. [22] Turns SR. An introduction to combustion. 3rd ed. Singapore: McGraw-Hill; 2012. p. 460e75. [23] Lee MC, Yoon J, Joo S, Yoon J, Kim M, Yoon Y. Investigation into the cause of high multi-mode combustion instability of H2/CO/CH4 syngas in a partially premixed gas turbine model combustor. Proc Combust Inst 2014. http://dx.doi.org/ 10.1016/j.proci.2014.07.013). in press. [24] Lefebvre AH, Ballal DR. Gas turbine combustion. 3rd ed. Boca Raton: Taylor & Francis; 2010. p. 364e6. [25] Shih W, Lee JG, Santavicca DA. Stability and emissions characteristics of a lean premixed gas turbine combustor. Proc Combust Inst 1996;26:2771e8. [26] Kim J, Yoon Y, Park CW, Hahn JW. The characteristic modes and structures of bluff-body stabilized flames in supersonic coflow air. Int J Aero Space Sci 2012;13(3):386e97. [27] Lieuwen TC, Yang V, Yetter R. Synthesis gas combustion. Boca Raton: Taylor & Francis; 2010. p. 273e8. [28] Coward HF, Jones GW. In: Limits of flammability of gases and vapors, 503; 1952. p. 5e9. U.S. Bureau of Mines Bulletin.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
14
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
[29] Chomiak J, Longwell JP, Sarofim AF. Combustion of low calorific value gases: problems and prospects. Prog Energy Combust Sci 1996;26:2771e8. [30] Warnatz J, Maas U, Dibble RW. Combustion. 4th ed. Berlin: Springer; 2006. p. 219e31. [31] Yoshimura T, McDonell V, Samuelsen S. Evaluation of hydrogen addition to natural gas on the stability and emissions behavior of a model gas turbine combustor. ASME Turbo Expo 2005; GT2005e68785. [32] LaViolette M, Perez R. On the prediction of pollutant emission indices from gas turbine combustion chambers. ASME Turbo Expo 2012; GT2012e70038. [33] LaViolette M, Strawson M. On the prediction of nitrogen oxides from gas turbine combustion chambers using neural networks. ASME Turbo Expo 2008; GT2008e50566.
[34] Lieuwen T, Torres H, Johnson C, Zinn BT. A mechanism of combustion instability in lean premixed gas turbine combustors. J Eng Gas Turbine Power 2001;123:182e9. [35] Seo S. Combustion instability mechanism of a lean premixed gas turbine combustor. J Mech Sci Technol 2003;17(6):906e13. [36] Kulsheimer C, Buchner H. Combustion dynamics of turbulent swirling flames. Combust Flame 2002;131:70e84. [37] Schadow KC, Gutmark E, Parr TP, Parr DM, Wilson KJ, Crump JE. Large-scale coherent structures as drivers of combustion instability. Combust Sci Technol 1989;64:167e86. [38] Lieuwen T, Yang V, Yetter R. Synthetic gas combustion. Florida: CRC Press; 2010. p. 261e88.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications Min Chul Lee a,1, Jisu Yoon b, Seongpil Joo b, Youngbin Yoon b,* a
Department of Safety Engineering, College of Engineering, Incheon National University, Academy-ro 119, Yeonsugu, Incheon 406-772, Republic of Korea b Department of Mechanical and Aeronautic Engineering, Seoul National University, Gwanack-Gu, Seoul 151-742, Republic of Korea
article info
abstract
Article history:
This paper describes the gas turbine combustion characteristics of coal-derived synthetic
Received 16 December 2014
gas (syngas), particularly for the syngases at the Taean IGCC plant in Korea and the Bug-
Received in revised form
genum IGCC plant in the Netherlands. To evaluate the combustion performance of these
13 June 2015
syngases, we conducted combustion tests with elevated temperature and ambient pres-
Accepted 16 June 2015
sure in a GE7EA model combustor. We observed flame stability, dynamic pressure char-
Available online xxx
acteristics, NOx and CO emissions, temperature in the combustion chamber, and flame structures while varying the heat input and diluent integration ratio. Without dilution,
Keywords:
Buggenum's stable regime is larger than that of Taean, since a higher hydrogen content
Integrated gasification combined
causes sustained flames, even at a very high flame stretch rate. However, when consid-
cycle
ering nitrogen dilution, Taean's syngas has a more stable burn than that of Buggenum,
Combustion performance test
since an increase in nitrogen at Buggenum has negatively affected flame stability. From the
Synthetic gas combustion
results of NOx/CO emissions and combustion efficiency, we report that both syngases are
Flame stability
sufficient to control NOx emissions at under 5 ppm with almost complete and stable
Combustion instability
combustion; however, quantitatively, the effect of nitrogen dilution is slightly different at
NOx emission
Taean and Buggenum due to slight differences in the H2/CO ratio and diluent heat capacity. All the tested results and conclusions drawn are considered for optimal operation and trouble shooting at the Taean IGCC plant, which is scheduled to be complete toward the end of 2016. Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ82 2 880 7396; fax: þ82 2 872 8032. E-mail addresses: [email protected] (M.C. Lee), [email protected] (Y. Yoon). 1 Tel.: þ82 32 835 8295; fax: þ82 32 835 0779. http://dx.doi.org/10.1016/j.ijhydene.2015.06.086 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Abbreviations C cp,i CCS Dh Dswirl_in Dswirl_out DRN2 hc IRN2 FLmix FLi IGCC LHV LFL m_ i m_ air m P3 P0ð3Þ; rms PIV ppm f Rair Re r Sn SCGP T4 T1 Tad t T3 q Ujet UHC Vpz xi [X]
Sonic velocity in the air at the temperature of T Specific heat at constant pressure of i-gas Carbon capture and storage Hydraulic diameter Inner diameter of swirler Outer diameter of swirler Dilution ratio of nitrogen Combustion efficiency Integration ratio of nitrogen Flammable limit of mixture Flammability limits of binary sub-mixtures Integrated gasification combined cycle Lower heating value Lower flammable limit Mass flow rate of i-gas Mass flow rate of air Dynamic viscosity Combustion chamber inlet pressure Root mean square of the dynamic pressure at the dump plane Particle image velocimetry Parts per million Equivalence ratio Ideal gas constant for air Reynolds number Density Swirl number Shell coal gasification process Combustor liner temperature Air inlet temperature Adiabatic flame temperature Residence time Combustion chamber inlet temperature Swirl vane angle Jet velocity Unburned hydro carbon Volume of the primary zone Volume percentage of a sub-mixture in total fuel Molar concentration of X species
Introduction Background IGCC is a potential technology for addressing the growing concern regarding global warming, urban atmospheric pollution, and depletion of energy resources. IGCC uses abundant low rank coal or renewable fuels such as biomass and agricultural or municipal waste, and offers environmentally clean benefits such as low NOx combustion, syngas desulfurization, and CCS capabilities. Although this promising technology has been developed and demonstrated since 1990 at numerous sites (e.g., Wabashi River and Tampa in the US, Puertollano in Spain, Buggenum in the Netherlands and Nakoso in Japan),
low plant availability in initial operation years is considered a serious issue. More precisely, unplanned outages due to fuel nozzle and liner burn out, combustion instability, and compressor surging in gas turbine units are the most critical sources of low availability [1e3]. Since we expect many outages in gas turbine combustors at new IGCC plants, due to the conditions at existing IGCC plants, a more detailed understanding of syngas combustion is necessary, and advanced syngas turbine technology should be adapted to improve IGCC production. The first Korean IGCC project (power output: 300 MWe; plant site: Taean) was launched in 2006 and is expected to be complete toward the end of 2015. For better operational availability and combustion optimization at this plant, we tested Taean syngas in a model GE7EA gas turbine combustor in this study to quantify potential application problems and evaluate its combustion performance as a gas turbine fuel.
Prior research on syngas turbine combustion Due to the growing interest in IGCC during the last two decades, many researchers have studied syngas combustion. These syngas combustion studies are divisible into two categories: fundamental research on combustion characteristicsdsuch as laminar burning velocity [4,5], ignition [6,7], flashback [8], blowout [9], flammable limit [10], and emissions [11e13]dand advanced research using gas turbine-like combustors under gas turbine-similar conditions [14e20]. As researchers in the first category, Natarajan et al. (2007) examined the effect of CO2 dilution, preheating, and pressure on the laminar flame speed of H2/CO mixtures by using two measurement approaches: flame area images of a conical Bunsen flame and velocity profile measurements of a onedimensional stagnation flame [4]. Walton et al. provided an experimental data set on the ignition delay time of hydrogen and carbon monoxide syngases under high pressure and high temperature conditions [6]. From the regression analysis of their data set, they obtained the best-fit correlation between ignition delay time and experimental variables. Furthermore, Lieuwen et al. investigated the impact of syngas composition on four combustor-operability issuesdblowout, flashback, combustion instability, and autoignitiondby analyzing calculations obtained from the CHEMKIN program [9]. Giles et al. investigated the NOx emission characteristics of syngas mixtures with airstream dilution of H2O, CO2, and N2 in a counterflow diffusion flame. They found that the NOx reduction effect of dilution was H2O > CO2 > N2 [12]. Such scientific foundations have provided useful insights and information to use when analyzing results from advanced studies, which enables the development of rather economical approaches to understand combustion phenomena in syngas combustors before conducting complicated tests on full-scale engines. Researchers in the second category have conducted studies, occasionally at great expense, to evaluate the combustion performance of syngas combustors and investigate syngas combustion characteristics in simulated gas turbine environments. Littlejohn et al. conducted combustion tests on a prototype injector of Taurus 70 engine for various syngas compositions and determined the lean blowout limits, loglinear dependency of NOx, and the turbulent flame speed
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
3
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
correlation using PIV data [14]. Dodo et al. investigated the performance of a multiple-injection, dry low NOx combustor with hydrogen-rich syngas fuels and reported the effectiveness of the convex perforated plate swirler on enhanced cooling of the combustor liner and suppression of combustion oscillation and NOx [15]. Through combustion tests of a can combustor of heavy-duty gas turbines, Hasegawa et al. reported many useful results related to NOx/CO emissions, pattern factors, combustion efficiency, and thermal efficiency [16e18]. Lee et al. also tested a can combustor for a GE7EA gas turbine combustor and reported the combustion performance characteristics of synthetic natural gas and H2/CO syngas with N2, CO2, and steam dilution [19e21]. They found that a 40% dilution of N2, CO2, and steam reduced NOx emission by 79%, 88%, and 95%, respectively, and that the NOx reduction per unit power is logarithmically related to the diluent's heat capacity. Although the abovementioned practical studies achieved technical progress in syngas turbine combustion, many unsolved problems remain due to the highly nonlinear behavior of combustion characteristics according to syngas composition, combustor geometry, and test conditions. In addition, our knowledge of syngas combustion in commercial gas turbines is mostly confined to commercial gas turbine makers, such as GE and Siemens, and operators of existing plants. Motivated by this consideration, we tested the Taean syngas to obtain a broad understanding of combustion characteristics under conditions relevant to a gas turbine. In particular, in this paper, we discuss the results from the combustion performance test in terms of flame stability, thermo-acoustic combustion instability, NOx and CO emissions, combustion efficiency, and flame structure. Furthermore, we compare this with the Buggenum syngas, which uses the same gasification technology as TaeandSCGP. The ultimate goal of this study is to provide appropriate solutions to syngas combustion issues and improve the efficiency and reliability of the Taean IGCC plant through an understanding of syngas combustion characteristics.
Test method The gas turbine combustion test facility and operating conditions For the purposes of this study, an atmospheric pressure combustion test facility for 60 kW-scale gas turbines was installed, as shown in Fig. 1. This facility consisted of an air compressor, air heater, a cooling and combustion air feed line, an atmospheric pressure combustor, an external stack with a silencer, and a fuel supplying system. Coal derived syngases are mainly composed of H2, CO, N2, CO2 and steam, and last three gases are incombustible diluents of which heating value is zero. Since these diluents are almost interchangeable in views of NOx/CO emission and combustion instability when the diluent's heat capacity is same [21]. Thus, only N2, the largest content in syngas, has been selected as a representative diluent to simplify the test. The fuel supply system can control the flow rate of H2, CO, and N2, respectively and both the fuel and air were processed to meet the choking condition immediately upstream of the combustor to prevent the perturbation of air and fuel in the combustion chamber by blocking the influence of pressure variation on the upstream flow. As shown in Table 1, the tests were conducted at a slightly elevated pressure, since a water-cooled plug nozzle blocked Table 1 e Experimental conditions. Items
Unit
Combustion chamber pressure Air inlet temperature Mean mixture jet velocity at combustor inlet Reynolds number Swirl number Heat input Overall equivalence ratio Nitrogen dilution ratio
Air compressor Regulator
Value
Bara C m/s
1.1e1.4 200 41.2e74.9
e e kWth L/L %
28,398e51,600 0.832 30e60 0.381e0.766 0e150
Flow FCV meter
Dryer
MFC
Filter
H2
Air heater MFC
Cooling air
CO
Combustion air
Regulator
Silencer Stack
N2 for purge Regulator
N2
MFC
GE7EA combustor
Gas sampling
Inline static mixer Fig. 1 e Schematic view of gas turbine combustion test facility.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
To offer a simpler description, the DRN2 and IRN2 are defined as shown in the following equations:
90% of the combustion outlet area to form an acoustic boundary for the combustion chamber outlet. The inlet temperature of combustion air was controlled at 200 C to meet conditions relevant for a gas turbine. According to the fuel and airflow rate, the average jet velocity of the fuel-air mixture at the combustor inlet varied from 41.2 m/s to 74.9 m/s, which are values that are comparable with those of a real engine. The Reynolds number, calculated by the following equation, was maintained between 28,398 and 51,600: Re ¼
rUjet Dh ; m
DRN2 ¼
IRN2 ¼
flow rate of N2 ½slpm : N2 equivalent dilution ratio total fuel flow rate ½slpm
where the N2 equivalent dilution ratio can be determined from Equations (2 and 3) by obtaining the m_ N2;eq of each syngases and substituting it into the flow rate of N2. DRN2 indicates the amount of nitrogen contained in the fuel compared to the combustible H2/CO gas, and IRN2 indicates the amount of nitrogen content in the fuel compared to the composition of the Taean or Buggenum syngases. That is, 100% of IRN2 is 56.4 mol % and 65.4 mol% nitrogen in the fuel at Taean and Buggenum, respectively. This represents the conditions under which all nitrogen produced from an air separation unit is supplied to the gas turbine combustor and, thus, when an IGCC plant is operating with the maximum integration of nitrogen.
(1)
m_ CO2 cp;CO2 þ m_ steam cp;steam þ m_ Ar cp;Ar ; cp;N2
(3)
(4)
where r, Ujet , Dh, and m are density of mixture, jet velocity of mixture, hydraulic diameter at combustor inlet, and dynamic viscosity, respectively. The turbulent Re was 45,000e81,000, assuming a relative turbulence intensity of 10% and an integral scale of 1/10 in the chamber diameter. Although chemical reaction times vary significantly according to the fuel composition and the amount of nitrogen dilution, the operational points in this research lie between the wrinkled laminar-flame regime and the flame in-eddies regime in the € hler-Turbulence Reynolds chart, as is true for other gas Damko turbines and internal combustion engines [22]. Fuel was supplied from each bottle of feedstock gases while keeping fuel composition constant and varying heating values from 30 kW to 60 kW, the span of which is 5 kW. Fuel gases were mixed well through the inline static mixer and supplied to the combustor. As previously described in the explanation on the fuel supply system, only nitrogen, instead of carbon dioxide and steam, was used as a diluent because the effect of dilution on combustion performance is almost the same when diluents have the same heat capacity [21]. Thus, the equivalent mass flow rate of nitrogen is determined from the following equation: m_ N2 ;eq ¼ m_ N2 þ
flow rate of N2 ½slpm : sum of flow rate of combustible gases ½slpm
Model gas turbine combustor To examine the combustion characteristics of the syngases, a 1/3 scaled-down one-can dump combustor of a GE7EA gas turbine was designed and fabricated as a test bed combustor. Fig. 2 shows the schematic of the combustion test rig. The combustion chamber consisted of two parts. The first part was made from optically accessible quartz, which was cooled by a high-pressure injection of the same amount air as the combustion air, and the second part was made from steel. To meet gas turbine relevant conditions, combustion air at 356 ± 5 C controlled by an mass flow controller (MFC, Bronkhorst F206BI, uncertainty ¼ ±0.8%) was supplied to the flame through a central annular swirling nozzle (swirl number ¼ 0.832, 14 swirl channels, i.d. ¼ 25 mm, o.d. ¼ 40 mm). Tests were conducted at a slightly aviated pressure (1.1e1.4 bar), since a 90% area of the combustor outlet was blocked by a water-cooled plug nozzle to form an acoustic boundary at the combustor
(2)
where m_ i and cp,i are mass flow rate and specific heat at constant pressure of i-gas.
Static pressure sensor Dynamic pressure sensor Temperature sensor T(4)
T(1) P
(1)
P
Burner head (2)
Fuel nozzle
P
(4)
P SP(2)
(5)
P
(6)
P
T(5) (7)
P
(8)
P SP(2)
(9)
P
(10)
P
(11)
T(2)
P
(3)
SP(1)
Ignitor T(3)
180 mm (Combustor liner - quartz part)
1230 mm (Combustor liner - steel part)
Plug nozzle
Fig. 2 e Schematic diagram of the gas turbine combustor and test rig with measurements. Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
5
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
outlet. The detailed description on combustor used for this study is given in Ref. [23].
Table 2 e Syngas composition at Taean and Buggenum [3].
Measurements
Syngas composition
As shown in Fig. 2, eleven piezoelectric dynamic pressure sensors (PCB 102A05), three static pressure sensors, and five ktype thermocouples were installed at the combustor and air inlet line to measure the dynamic and static pressure and temperature of each location. To minimize the measurement error of temperature at combustor due to radiative heat transfer from thermocouple to combustor wall, temperature has been measure at the thermally stabilized condition. However, thermocouples are not radiation-shielded so an error around 5% of measured temperature can be introduced. Exhaust gas was sampled at the combustor outlet, and the concentrations of NOx, CO, O2, CO2, and UHC were measured by a TESTO 360 gas analyzer. To measure the NOx and CO emissions, the exhaust gas was sampled at the exhaust end of the combustor and analyzed using a TESTO 360 gas analyzer. The measured results for emissions were corrected to 15% oxygen on a dry basis using the following equations [24]
H2 CO CH4 N2 CO2 H2O Ar H2S Sum Higher heating value Lower heating value H2/CO Mole fraction of combustible gas Lower flammable limit in air Upper flammable limit in air N2 equivalent dilution ratio Upper Wobbe index Lower Wobbe index
ðmeasured NOx Þ ðNOx at 15% O2 Þ ¼ 29:79 ) ( 4:76 ð2 ðmeasured O2 ÞÞ 1 1 4:76 ðmeasured O2 Þ ðmeasured CO Þ ðCO at 15% O2 Þ ¼ 29:79 ) ( 4:76 ð2 ðmeasured O2 ÞÞ 1 1 4:76 ðmeasured O2 Þ
Unit Taean Buggenum Difference [A] [B] [(A B)/A] mol% 12.9 mol% 31.5 mol% 0.1 mol% 50.8 mol% 0.4 mol% 4.2 mol% 0.4 mol% 0.0 mol% 100.0 MJ/Nm3 5.60 5.35 MJ/Nm3 % 40.6 mol% 44.3
12.3 24.8 0.0 42.4 0.8 19.1 0.6 0.0 100.0 4.70 4.46 49.6 37.1
3.54% 21.12% 0% 17.09% 114.57% 341.86% 37.09% 0% 0% 16.07% 16.64% 22.47% 16.06%
%
19.1
21.2
10.99%
%
73.3
71.7
2.18%
%
56.4
65.4
15.96%
MJ/Nm3 MJ/Nm3
10.02 9.57
9.44 8.95
5.79% 6.48%
(5)
(6)
To investigate flame structure, an OH*-chemiluminescence image of flame was obtained using a PI-MAX Gen II ICCD camera with two optical filters (UG11, WG305) and a UVNikkor 105 mm lens. Because OH*-chemiluminescence images are line-of-sight integrated images, a three-point Abel deconvolution scheme was used to extract two-dimensional information from the line-of-sight integrated images.
Results and discussion Combustion tests on both fuel gases were conducted with respect to the heat input and DRN2, and the results were compared to understand the effect of slight differences in syngas composition on combustion performance. We also compared the results to predict various combustion phenomena that occur when Taean or Buggenum syngas fuels a GE7EA gas turbine. Table 2 outlines the fuel composition of Taean and Buggenum, which were determined from the reports of Taean and Buggenum IGCC plants [3]. The main features of the Taean syngas compared with the Buggenum syngas are a lower H2/CO ratio, lower diluents in fuel due to low moisture, and thus higher heat content per unit volume.
Flame stability map In the present study, we classified the stability regime according to the flame shape and intensity of pressure
fluctuation, as reported by many flame stability map researchers [25e27]. Fig. 3 describes the criteria for classification on the flame stability map. Regime I is a V-shaped flame; this flame shape occurs due to swirling flow and outwardly diverging recirculation flow toward the dump plane when the flame root is attached to the fuel nozzle. Regime II is an Mshaped flame; as opposed to regime I, this type of flame interacts with a vortex flow and burns close to the dump plane and quartz liner. In regime IV, a long cone-shaped flame occurs, and it seems like a long tornado due to the swirling flow. In regime III, flames of regimes II and IV occur alternately, and this flame oscillation results in large pressure fluctuations. In regimes V and VI, lift off and blowout phenomena occur, respectively. Fig. 4(a) shows the flame stability map of Taean and Buggenum syngas according to DRN2. For both types of fuel, flame changes from strong burning regime I through II, III, IV, and V to regime VI as the DRN2 increased and the heat input decreased. Regime I occurred at just over 40 kW of heat input with a DRN2 of under 100%. As the nitrogen dilution increased, more enhanced recirculation flow occurred due to the high fuel-air mixture momentum, which resulted in the flame's transition from regime I to regime II. Further nitrogen additions in syngas slowed down the combustion reaction, and consequently lowered the burning velocity. Moreover, since a nitrogen addition increases the mean mixture velocity, the flame was extinguished locally, being detached from the dump plane and moving downstream when the mean mixture velocity exceeded the turbulent burning velocity. This movement downstream resulted in a regime shift from II to VI, passing through a narrow region of III, in which the flame oscillated rather unstably with an alteration motion between detachment and attachment to the dump plane. Next, flame lifting (regime V) and lean blowout (VI) phenomena occurred
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Regime
Flame images
Description
V-shaped flame I
(Flame does not interact with outer recirculation flow)
M-shaped flame II
(Flame interacts with outer recirculation flow)
Oscillating flame (Flame oscillates alternately between M-shaped flame and
III
long cone-shaped flame)
Long cone-shaped flame (Flame does not interact with outer recirculation flow and
VI
seems like a tornado due to swirling flow)
Lift-off
V
VI
Blowout
No Flame
Fig. 3 e Criteria for classification on a flame stability map.
sequentially with decreasing heat input and increasing nitrogen dilution. Combustion instability did not occur for all tested conditions without regime III (detailed results and discussions on the combustion instability characteristics are in section 5 of chapter 3). Therefore, when considering only the flame stability map, we conclude that only regimes I and II are stable operating conditions for Taean and Buggenum syngas combustion, and the Taean syngas is more stable when compared to the Buggenum syngas since operating condition of Taean is located farther from the unstable regime III than that of Buggenum. In the meantime, we redrew the flame stability map according to the nitrogen mole fraction in fuel, as depicted in Fig. 4(b). This figure illustrates several important points missing from Fig. 4(a). First, in the low load condition, excessive nitrogen dilution (exceeding 30 mol%) can cause unstable
combustion. Second, all borderlines (except the border between regimes I and II) asymptotically approach the range between 77 mol% and 82 mol%. Third, the flames of both fuels cannot exist when the nitrogen content is over 82 mol% in the fuel. To understand why a flame blows out over a certain limit, we estimated the LFL of the H2/CO/N2 mixture on the basis of the Le Chatelier rule [28,29]: FLmix ¼
x1 FL1
100 ; x3 x2 þ FL þ FL þ / 2 3
(7)
where FLmix is measures the flammability limit of the mixture in vol%, FLi is the flammability limit of the binary submixtures comprised of one combustible gas and one inflammable gas, and xi is the volume percentage of a sub-mixture in P total fuel ( xi ¼1). Assuming that the sub-mixtures are H2 þ N2 and CO þ N2, we calculated the maximum LFL of
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
7
Fig. 5 e Temperature characteristics of Taean and Buggenum syngases with respect to the nitrogen integration ratio.
Fig. 4 e Flame stability map of Taean and Buggenum syngases.
Taean and Buggenum syngases to be 83.9 vol% and 84.5 vol%, respectively. These value lines are plotted in Fig. 4(b), and the blowout limit occurred because both fuels approach the LFL lines of syngas but cannot exceed them. It must be noted that even though a high swirl stabilizes a flame almost until the LFL, the flame cannot go beyond the maximum LFL of syngas. Inversely, the blowout limit line of any mixture can be estimated by obtaining the maximum LFL of a mixture.
Temperature characteristics Fig. 5 presents measurements of T1 and T4 with respect to IRN2. With these temperatures, the Tad, calculated via CANTERA code with the GRI 3.0 mechanism, is plotted. For all test
conditions, the T1 is maintained at 200 ± 3 C, which indicates that the air preheater controls T1 well and that the test's accuracy in terms of temperature is over 98.5%. As expected, T4 and Tad show a linearly increasing tendency as heat input increases because the flame temperature increases following the rise in heat input. As the IRN2 increases, the Tad decreased almost linearly since the addition of inflammable nitrogen increased the total heat capacity in the fuel and, thus, the heat demand to raise the flame temperature. The nitrogen-induced flame cooling effect at Buggenum is stronger than that at Taean, since the Buggenum syngas has a higher nitrogen mole fraction in the fuel, as shown in Table 2. This fact implies that the Buggenum IGCC plant expends 16% more nitrogen to control NOx emissions, thereby resulting in a 8.2% larger decrease in Tad at 60 kW with 100% DR (Tad of Taean has decreased from 2066 C to 1627 C whereas Tad of Buggenum has decreased from 2071 C to 1507 C). We observed similar characteristics in T4 for the same reasons. However, we observed a sudden drop in T4 at over 50% of IRN2. Flame position and structure largely affected T4, so the stability mode change from regime II to regime IV via regime III resulted in the change in T4. For example, for the Taean syngas at 30 kW, we observed a sudden drop between 39% and 50% of IRN2, and this condition corresponds to the 47.3% and 86.4% of DRN2 in regimes II and IV, respectively, according to Fig. 4(a). On the other hand, for the Buggenum syngas at 30 kW, we observed a sudden temperature drop between 50% and 100% of IRN2, and this condition corresponds to the 88.2% and 176.3% of DRN2 in regimes III and IV, respectively. Thus, it can be concluded that flame movement downstream is the main cause of a sudden drop in combustor liner temperature. From this result, it can be concluded that nitrogen for NOx reduction should be carefully controlled in a lower load (under 40 kW) since flames are easily destabilized. Consequently, the flame might lift off and then damage downstream portions of the combustor, such as turbine blades and vanes.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
8
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
NOx and CO emission characteristics Fig. 6(a and b) depict the emission characteristics of NOx and CO with respect to IRN2. For both syngases, the concentration of NOx emissions showed an exponentially decreasing trend when adding more nitrogen to the fuel for all heat inputs. We attributed these results to the decrease in flame temperature via increased nitrogen integration and the extended Zeldovich NOx mechanism, which explains that NOx emissions exponentially correlate with flame temperature [30]. From the three Zeldovich reactions, assuming a quasi-steady state of nitrogen atoms, we obtained the following NO-formation equation d½NO ¼ 2k1 ½O½N2 ; dt
(8)
where the rate coefficient k1 ¼ 1.8∙1014 exp (318 kJ mol1/ (RT)) [cm3/mol-s], and [X] is the molar concentration of X species. Thus, NO can be minimized by decreasing [N2], [O], or k1 (i.e., by decreasing the flame temperature). It is noteworthy that increased N2 enhances NOx emissions in view of [N2] and simultaneously suppresses NOx in view of flame temperature. However, flame temperature dominates the NOx emission, and thus NOx decreases as N2 increases because k1 is insignificant at temperatures below 1700 K [31]. This feature is evident in Fig. 7, which illustrates the NOx emission dependence on Tad which has been already mentioned in chapter 3.2. For both syngases, NOx emissions become significant at Tad > 1450 C and increase exponentially as Tad increases. Furthermore, most of the test data fit well in the exponential asymptotic curve. This good fit implies that Tad mainly affects NOx, not the nitrogen mole fraction. Discordance and early rise of NOx in low heat input conditions might be attributed to the nature of a partially premixed flame, which forms a localized high temperature zone. NOx emissions of the Taean syngas were slightly higher than that of Buggenum by 2.09 ppm at 60 kW without N2 dilution. Since flame temperature is almost the same for that condition, residence time in flame can be considered as the main cause of the NOx emission difference. We calculated the residence time using the following equation, which is based on hot gas flow through the primary flame zone [32,33]: t¼
Vpz P3 ; m_ air Rair T3
Fig. 6 e NOx and CO emissions of Taean and Buggenum syngases.
(9)
where Vpz, P3, m_ air , Rair, and T3 are volume of the primary zone [m3], combustion chamber inlet pressure [Pa], mass flow rate of air [kg/s], ideal gas constant for air [J/kg-K], and combustion chamber inlet temperature [K], respectively. Herein, P3, m_ air , Rair, and T3 are the same for both syngases and only Vpz is a control variable of t, which is obtained from the timeaveraged Abel-deconvoluted OH*-chemiluminescence images in Fig. 8. The calculated Vpz of the Taean syngas was 5020 mm3, which was larger than that of Buggenum by 8.2%; consequently, the t of the Taean syngas was longer than that of Buggenum by 4%. Therefore, we can assert that an 8.2% longer t in a flame is a vital cause of increased NOx emissions, when comparing the Taean and Buggenum syngases. This volume difference might be due to the H2/CO ratio in the fuel. Since H2 burns more than 10 times faster than CO (see
Fig. 7 e NOx emissions of Taean and Buggenum syngases with respect to the adiabatic flame temperature.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
9
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Fig. 8 e Primary combustion zone of Taean and Buggenum syngases (left-half images of time-averaged Abel-deconvoluted OH*-chemiluminescence).
Table 3), the Taean syngas's flame propagation speed, which contains more CO, must be lower than that of the latter. Hence, this relatively slow reaction resulted in a larger Vpz and longer t. For both tested syngases, the concentration of CO was very high, reaching up to 500 ppm or more for lower loads with high nitrogen integration (over 50%), but the concentration decreased sharply for higher loads (down to 10 ppm or less), showing almost perfect combustion (combustion efficiency > 99.9%). Because the high emissions of CO stemmed from the low mixed mean adiabatic combustion temperature, the nitrogen injection (which quenches flame) can be a main source of high CO emissions for the low load. When increasing the load, sufficient main fuel supply can cause the flame temperature to rise, and thus CO emissions drop to a sufficiently lower level. Fig. 9 illustrates the relationship between CO emissions and flame temperature. CO is rapidly emitted
when Tad < 1230 C, and this spiking relationship between CO and flame temperature coincides with the results from Yoshimura et al. [31], who reported that CO emissions exponentially increase at Tad < 1500 K. By comparing the Buggenum and Taean syngases, we found that the Buggenum syngas is prone to emit more CO. For example, over 1000 ppm of CO is emitted when firing the Buggenum syngas at 40 kW with 150% IRN2. On the other hand, we detected no significant CO emission under the same conditions for the Taean syngas. This discrepancy can be explained by the fact that the Buggenum IGCC plant integrates more nitrogen to control NOx, as shown in Table 2, because the Tad at Buggenum is significantly lowered when IRN2 increases and the slope is steeper than that of the Taean syngas, as shown in Fig. 6(a). From the NOx and CO emission results, we conclude that emissions characteristics highly correlate with flame temperature, which relates to IRN2 as well as the H2/CO ratio.
Table 3 e Properties of hydrogen, carbon monoxide, and methane [19,20,33e35]. Property Chemical formula Boiling point Specific gravity Flammable limits in air Autoignition temperature Max. burning velocity Max. adiabatic flame temperature Stoichiometric air/fuel ratio Higher heating value Lower heating value Upper Wobbe index Lower Wobbe index
Unit
Hydrogen
Carbon monoxide
Methane
e C (vs. air) % C cm/s C kg/kg MJ/Nm3 MJ/Nm3 MJ/Nm3 MJ/Nm3
H2 252.8 0.07 4e75 500 289 2200 34.1 12.76 10.80 48.23 40.65
CO 191.5 0.97 12.5e74.2 609 19 2210 2.5 12.63 12.63 12.80 12.80
CH4 161.5 0.55 5.0e15.4 537 37 2000 16.9 39.75 35.82 53.28 47.91
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
10
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
lower flame temperature, and these emissions were suppressed to under 10 ppm when the heat input was over 45 kW.
Combustion oscillation characteristics
Fig. 9 e CO emissions of Taean and Buggenum syngases with respect to the adiabatic flame temperature.
However, even though an increased nitrogen molar concentration should enhance NOx formation, according to the extended Zeldovich mechanism, this effect was not significant in our study when compared to the effect of flame temperature. When considering only the H2/CO ratio, the Taean syngas emitted more NOx than the Buggenum syngas due to a larger t; these emissions were suppressed to under 3 ppm when IRN2 was over 50%. On the other hand, when increasing the IRN2, the Buggenum syngas emitted more CO because of a
Fig. 10 illustrates our investigation of combustion oscillations of both syngases with respect to IRN2. The combustion pulsation level was very low in all tests (i.e., there was almost zero combustion instability without dilution in the case of 35 kW with 100% IRN2). This low fluctuating level is much lower than the oscillation in a standard gas turbine combustor, and it is at least 5% under the control limit for oscillation in a standard turbine combustor. These results are comparable to the results of previous tests on H2 and CO combustion with and without dilution [20,21]. To understand why combustion instability did not occur, two mechanisms of particular significance in gas turbine combustion systems can be introduced: oscillation feedback among acoustic pressures, equivalence, and heat release [34,35] and flame-vortex interaction [36,37]. In the former mechanism, an acoustic pressure fluctuation in the combustor results in a velocity and acoustic pressure fluctuation at the fuel nozzle, which perturbs the fuel flow rate. This perturbation propagates to the flame, where it produces a heat release oscillation and, in turn, another pressure fluctuation. Here, heat release oscillation is the periodic repetition of extinction and re-ignition, and a syngas with considerable hydrogen content cannot be easily extinguished due to the large flammable limit and flame extinction
Fig. 10 e RMS values of pressure fluctuations with nitrogen integration ratio (Upper: Taean syngas; Lower: Buggenum syngas). Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
stretch rate of highly reactive hydrogen [38]. Thus, fluctuations in heat release are minimal and do not cause fluctuations in pressure or fuel flow rate. In the latter mechanism, large-scale, coherent vortical structures cause an oscillation in the rate of heat release. These structures are the result of flow separation from flame holders and rapid expansions, as well as vortex breakdown in swirling flows. However, in this study, neither a vortical flame structure at the outer recirculation zone nor flow separation near the fuel nozzle is observed (as shown in Fig. 12) since hydrogen (whose burning velocity is very high) causes the flame shape to be diffusive and short. In the case of 35 kW with 100% IRN2, we observed a large amplitude pressure fluctuation in both syngases. Thus, to analyze the oscillation frequency, we plotted Fig. 11 by conducting the Fast Fourier Transformation. The main frequency was 10 Hz, which was due to flame oscillation in regime III, and 300 Hz and higher frequencies represented fundamental longitudinal frequency and their harmonics. Therefore, only flame oscillation near flame lift-off conditions can be considered as a significant source of combustion instability for both syngases, and flames in regimes I and II were not a concern for thermo-acoustic instabilities.
Flame structures Fig. 12 illustrates the time-averaged OH*-chemiluminescence images (upper half) and their Abel-deconvoluted images (lower half). Since the flow is assumed to be axisymmetric, only half images are shown. In low heat input conditions, the flame is weak and broader, because the outer recirculating flow can easily affect a flame at a low equivalent ratio. When increasing the heat input, a flame grows larger and stretches downstream, and the primary combustion zone moves to the center due to increased momentum in the downstream direction. When increasing IRN2, the burning intensity in both syngases diminishes and the flame moves from upstream/center to downstream/outer wall of the combustion chamber, thereby increasing the combustion region. The high temperature combustion region in which the thermal NOx occurred
Fig. 11 e FFT of pressure fluctuations of Buggenum syngas at 35 kW with 100% nitrogen dilution.
11
contracted substantially, but the lower temperature combustion zone expanded. As previously described in subchapter 3.3, these flame structure results coincide with the results of NOx and CO emissions.
Conclusion We conducted combustion performance tests of the Taean and Buggenum syngases in a 1/3 scaled-down GE7EA industrial gas turbine combustor. We obtain the following conclusions from the test results: (1) We plotted the flame stability map by categorizing flames into six regimes according to the flame shape and pressure fluctuation. We observed stable combustion in regimes I and II, into which both Taean and Buggenum syngases fall. When considering only the H2/ CO ratio in the fuel, the Buggenum syngas is more stable. However, when considering all aspects, including IRN2, the Taean syngas is more stable. Furthermore, flames from both syngases cannot exist when the nitrogen level is over 82 mol%, and this content almost corresponds to the maximum LFL estimated by the Le Chatelier rule. Therefore, we concluded that the lean blowout limit of syngas is predictable by simply calculating maximum LFL, even in the high swirl stabilized flame. (2) Temperature characteristics with respect to IRN2 and heat input showed monotonic increasing or decreasing trends, except at certain unstable flame conditions. In these unstable conditions, we observed sudden drops and elevations in combustor temperature due to flame movement. This condition should be avoided since it can damage the combustor's liner and transition piece as well as diminish the lifetime of hot gas path components, including turbine blades and vanes. (3) NOx and CO emissions highly correlate with flame temperature, and the optimal operating conditions of syngas turbines is determined to be when 1230 C < Tad < 1500 C. Under the same flame temperature, the Taean syngas, which contains more CO in the fuel, emits more NOx than the Buggenum syngas does due to a larger t at the former's primary combustion zone. On the other hand, the Buggenum syngas, which contains more diluents, emits more CO because of a lower flame temperature. We verified that fuel-side nitrogen dilution is very effective in suppressing NOx emission by lowering the flame temperature, and NOx emission decreases to under 3 ppm when IRN2 is over 50%. However, operating with a low load and a high DRN2 caused problems with CO emission and flame stability. Thus, the dilution system of an IGCC plant should be operated with analysts cautiously monitoring the CO emission and flame stability for better fuel efficiency and plant reliability. (4) Combustion instability occurred only near regime III due to the flame attachment/detachment mechanism. Thus, operating conditions, including IRN2, should avoid this combustion instability. In the other regimes, we did
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
12
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
Fig. 12 e OH*-chemiluminescence images (upper) and Abel-deconvoluted images (lower) for Taean and Buggenum syngases.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
not observe combustion instability for two reasons: 1) lower heat release oscillation and its feedback to pressure fluctuation, and 2) flame shape, which does not generate vortical flow. (5) We used flame structure to analyze the test results of combustion performance, such as flame stability, NOx and CO emissions, combustion efficiency, combustion instability, and temperatures in the combustor. All tested data will be used to determine the optimal operating conditions at the Taean IGCC plant and analyze the plant's outage. (6) Since IGCC with CCS technology is the one of the key technology to solve the global warming problem, many recent demonstration projects such FutureGen in US, GreenGen of China, ZeroGen of Australia have been conducted in the world and Korea government is also planning to construct two more 300 MWe IGCC plants by 2020. However, competing with other new thermal power technology such as ultra-supercritical pulverized coal boiler and circulating fluidized bed technology in views of economics and efficiency and recent large fluctuations of oil and gas price can vary the IGCC perspective in thermal power plant market.
Acknowledgments This work was supported by the Development of 300 MWclass Korean IGCC demonstration plant technology of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant, funded by the Korean Ministry of Trade, Industry and Energy (201195101001C). This work was also supported by the Mid-career Researcher Program, via an NRF grant, funded by the MSIP (NRF-2010-0015100), contracted through the Institute of Advanced Aerospace Technology at Seoul National University.
references
[1] Hornick MJ, McDaniel JE. Tampa Electric Polk power station integrated gasification combined cycle project final technical report. U.S. DOE project 2002; DE-FC21e91MC27363. [2] Dowd RA. Wabash River coal gasification repowering project final technical report. U.S. DOE project 2000; DEFC21e92MC29310. [3] Phillips J. Integrated gasification combined cycle design consideration for high availability. EPRI Technical Report. 2007. p. 1012226. [4] Natarajan J, Lieuwen T, Seitzman J. Laminar flame speeds of H2/CO mixtures: effect of CO2 dilution, preheat temperature, and pressure. Comb flame 2007;151:104e19. [5] Hu E, Huang Z, He J, Jin C, Zheng J. Experimental and numerical study on laminar burning characteristics of premixed methane-hydrogen-air flame. Int J Hydrogen Energy 2009;34:4876e88. [6] Walton SM, He X, Zigler BT, Wooldridge MS. An experimental investigation of the ignition properties of hydrogen and carbon monoxide mixtures for syngas turbine applications. Proc Combust Inst 2007;31:3147e54.
13
[7] Kalitan DM, Petersen EL, Mertens JD, Crofton MW. Ignition of lean CO/H2/air mixtures at elevated pressures. Proc of ASME Turbo Expo. 2006. GT2006:90488. [8] Xu G, Tian Y, Song Q, Fang A, Cui Y, Yu B, et al. Flashback limit and mechanism of methane and syngas fuel. Proc of ASME Turbo Expo. 2006. GT2006:90521. [9] Lieuwen T, McDonell V, Petersen E, Santavicca D. Fuel flexibility influences on premixed combustor blowout, flashback, autoignition, and stability. J Eng Gas Turb Power 2008;130:011506. [10] Noble DR, Zhang Q, Shareef A, Tootle J, Meyers A, Lieuwen T. Syngas mixture composition effects upon flashback and blowout. Proc of ASME Turbo Expo. 2006. GT2006e90470. [11] Whitty KJ, Zhang HR, Eddings EG. Emissions from syngas combustion. Comb Sci Tech 2008;180:1117e36. [12] Giles DE, Som S, Aggarwal SK. NOx emission characteristics of counterflow syngas diffusion flames with airstream dilution. Fuel 2006;85:1729e42. [13] Konnov AA, Dyakov IV, Ruyck JD. Nitric oxide formation in premixed flames of H2 þ CO þ CO2 and air. Proc Combust Inst 2002;29:2171e7. [14] Littlejohn D, Cheng RK, Noble DR, Lieuwen T. Laboratory investigations of low-swirl injectors operating with syngases. J Eng Gas Turb Power 2010;132:011502. [15] Dodo S, Asai T, Koizumi H, Takahashi H, Yoshida S, Inoue H. Performance of a multiple-injection dry low NOx combustor with hydrogen-rich syngas fuels. J Eng Gas Turb Power 2013;135:011501. [16] Hasegawa T, Tamaru T. Gas turbine combustion technology reducing both fuel-NOx and thermal-NOx emissions for oxygen-blown IGCC with hot/dry synthetic gas cleanup. J Eng Gas Turb Power 2007;129:358e69. [17] Hasegawa T, Nakata T. A study of combustion characteristics of gasified coal fuel. Trans ASME 2001;123:22e32. [18] Nakata T, Sato M, Ninomiya T, Hasegawa T. A study on low NOx combustion in LBG-fueled 1500 C-class gas turbine. Trans ASME 1996;118:534e40. [19] Park S, Kim U, Lee MC, Kim S, Cha D. The effects and characteristics of hydrogen in SNG on gas turbine combustion using a diffusion type combustor. Int J Hydrogen Energy 2013;38:12847e55. [20] Lee MC, Seo SB, Chung JW, Kim SM, Joo YJ, Ahn DH. Gas turbine combustion performance test of hydrogen and carbon monoxide synthetic gas. Fuel 2010;89:1485e91. [21] Lee MC, Seo SB, Yoon J, Kim M, Yoon Y. Experimental study on the effect of N2, CO2, and steam dilution on the combustion performance of H2 and CO synthetic gas in an industrial gas turbine. Fuel 2012;102:431e8. [22] Turns SR. An introduction to combustion. 3rd ed. Singapore: McGraw-Hill; 2012. p. 460e75. [23] Lee MC, Yoon J, Joo S, Yoon J, Kim M, Yoon Y. Investigation into the cause of high multi-mode combustion instability of H2/CO/CH4 syngas in a partially premixed gas turbine model combustor. Proc Combust Inst 2014. http://dx.doi.org/ 10.1016/j.proci.2014.07.013). in press. [24] Lefebvre AH, Ballal DR. Gas turbine combustion. 3rd ed. Boca Raton: Taylor & Francis; 2010. p. 364e6. [25] Shih W, Lee JG, Santavicca DA. Stability and emissions characteristics of a lean premixed gas turbine combustor. Proc Combust Inst 1996;26:2771e8. [26] Kim J, Yoon Y, Park CW, Hahn JW. The characteristic modes and structures of bluff-body stabilized flames in supersonic coflow air. Int J Aero Space Sci 2012;13(3):386e97. [27] Lieuwen TC, Yang V, Yetter R. Synthesis gas combustion. Boca Raton: Taylor & Francis; 2010. p. 273e8. [28] Coward HF, Jones GW. In: Limits of flammability of gases and vapors, 503; 1952. p. 5e9. U.S. Bureau of Mines Bulletin.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086
14
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 5 ) 1 e1 4
[29] Chomiak J, Longwell JP, Sarofim AF. Combustion of low calorific value gases: problems and prospects. Prog Energy Combust Sci 1996;26:2771e8. [30] Warnatz J, Maas U, Dibble RW. Combustion. 4th ed. Berlin: Springer; 2006. p. 219e31. [31] Yoshimura T, McDonell V, Samuelsen S. Evaluation of hydrogen addition to natural gas on the stability and emissions behavior of a model gas turbine combustor. ASME Turbo Expo 2005; GT2005e68785. [32] LaViolette M, Perez R. On the prediction of pollutant emission indices from gas turbine combustion chambers. ASME Turbo Expo 2012; GT2012e70038. [33] LaViolette M, Strawson M. On the prediction of nitrogen oxides from gas turbine combustion chambers using neural networks. ASME Turbo Expo 2008; GT2008e50566.
[34] Lieuwen T, Torres H, Johnson C, Zinn BT. A mechanism of combustion instability in lean premixed gas turbine combustors. J Eng Gas Turbine Power 2001;123:182e9. [35] Seo S. Combustion instability mechanism of a lean premixed gas turbine combustor. J Mech Sci Technol 2003;17(6):906e13. [36] Kulsheimer C, Buchner H. Combustion dynamics of turbulent swirling flames. Combust Flame 2002;131:70e84. [37] Schadow KC, Gutmark E, Parr TP, Parr DM, Wilson KJ, Crump JE. Large-scale coherent structures as drivers of combustion instability. Combust Sci Technol 1989;64:167e86. [38] Lieuwen T, Yang V, Yetter R. Synthetic gas combustion. Florida: CRC Press; 2010. p. 261e88.
Please cite this article in press as: Lee MC, et al., Gas turbine combustion characteristics of H2/CO synthetic gas for coal integrated gasification combined cycle applications, International Journal of Hydrogen Energy (2015), http://dx.doi.org/10.1016/ j.ijhydene.2015.06.086