Dynamic simulation of an oxygen mixed conducting membrane-based gas turbine power cycle for CO2 capture

Available online at www.sciencedirect.com

Energy Procedia

Energy Energy ProcediaProcedia 00 (2008) 000–000 1 (2009) 431–438

www.elsevier.com/locate/procedia www.elsevier.com/locate/XXX

GHGT-9

Dynamic Simulation of an Oxygen Mixed Conducting Membrane-based Gas Turbine Power Cycle for CO2 Capture Konrad Eichhorn Colombo, Olav Bolland* Department of Energy and Process Engineering, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway Elsevier use only: Received date here; revised date here; accepted date here

Abstract This paper investigates the transient behaviour of an oxygen mixed conducting membrane (OMCM)-based gas turbine (GT) power plant. Several operation and material constraints limit the operability of the power plant. For part-load operation two strategies are analysed: (i) reduction in mass flow of air to the GT in conjunction with reduced fuel supply to the afterburner while keeping the turbine exit temperature (TET) constant (TET control strategy), and (ii) reduction of fuel supply to the afterburner at constant air supply to the GT while the TET is allowed to vary (turbine inlet temperature (TIT) control strategy). Simulation reveals that this GT power plant shows rather slow dynamics because of the recirculation of large amount of gas. The OMCM-based GT power plant is compared to a simple GT power plant with respect to design, off-design as well as transient behaviour during load reduction. Information about controlled and manipulated variables in the GT power plant is given for the development of control strategy. © Ltd. All All rights rights reserved. reserved c 2008

2009 Elsevier Elsevier Ltd. Keywords: Oxygen Mixed Conducting Membrane; Dynamic Simulation; Gas Turbine Power Plant; CO2 Capture; Modelling; Reactor; Monolith; Part-Load; Control

1. Introduction Based on scientific analysis, it is generally accepted that there is a link between climate change and the emission of greenhouse gases such as CO2. The development of CO2 capture and storage technologies has therefore received increased attention worldwide. The oxy-combustion gas turbine (GT) power plant concept analysed in this paper is one such technology where combustion takes place in a nitrogen-free atmosphere provided that natural gas is free of such impurities [1]. Net efficiencies over 50 per cent has been reported when connecting this GT power plant to a steam cycle [2]. However, replacing mature gas turbine power plant technology by novel CO2 capture concepts needs to be proven with regard to operation reliability and cost of critical process units in addition to overall power plant performance. The oxy-combustion power plant incorporates an oxygen mixed conducting membrane (OMCM) based on perovskite-type material. In general, these oxides require elevated temperatures, typically in excess of 700°C [3], and an oxygen partial pressure gradient to transport the oxygen through the membrane. In the absence of cracks and leaks the OMCM is permeable for oxygen only. Such OMCMs applied for air separation lead to several additional constraints in the GT power plant. Operation beyond these limits results for instance in large thermo-chemical-mechanical stresses which may lead to irreversible damage [4]. Further constraints are set by the natural gas combustion in a highly diluted H2O-CO2 atmosphere with low oxygen content. From an operational point of view, this GT power plant is very challenging. In what follows, the OMCM-based GT power plant is briefly explained. In this respect, operation constraints of individual power plant units are emphasised. The design case is discussed and compared to a simple GT power plant followed by the analysis of both power plants at part-load. Afterwards, the transient behaviour of the OMCM-based GT power plant during load reduction is analysed and again compared to that of the simple GT power plant. For part load operation, two different control strategies are studied. * Corresponding author. Tel.: +47-73-59-1604; fax: +47-73-59-8390. E-mail address: [email protected]

doi:10.1016/j.egypro.2009.01.058

2

432

Konrad Eichhorn Colombo, OlavO.Bolland/ Procedia 00 (2008) 000–000 K.E. Colombo, BollandEnergy / Energy Procedia 1 (2009) 431–438

2. Gas Turbine Power Plant 2.1. Oxygen Mixed Conducting Membrane-based Gas Turbine Power Plant The OMCM-based GT power plant is shown in Fig.1. The combustor of a simple GT power plant, which is shown in Fig.2, is replaced by the OMCM reactor. The OMCM reactor comprises an OMCM, ceramic heat exchangers (HX) [5], combustor, and sub-sonic ejector. The ejector is applied for fuel injection into the OMCM reactor. Adding steam to the fuel provides the required force while keeping the temperature of the combustor limited. Each ejector is connected to a number of 12 monolith assemblies, consisting of high-temperature HX (HTHX), OMCM, and low-temperature HX (LTHX). The fluid distribution is accomplished by the flow manifolds which are illustrated in Fig.1. The power output of the OMCM-based GT power plant in design is 19.3 MW. That power output results in a total number of 576 monolith assemblies, and consequently 48 ejectors and combustors. A certain fraction of the sweep gas after combustion is split and cooled by the remainder air in the bleed-gas HX (BHX). Additional power can be produced by feeding this gas stream, mainly consisting of H2O and CO2, to a steam cycle (not shown in Fig. 1) before flash condensation, purification, and compression (not shown in Fig. 1). The exhaust gas from the GT can also be utilised for power generation in the steam cycle. Pipe models have been adopted to account for additional pressure and heat losses in the power plant.

Fig. 1: Oxygen mixed conducting membrane-based gas turbine power plant with values for key parameters in design. The OMCM, HXs, pipe and ejector models are based on conservation balances for energy, species, and total mass. Overall pressure drops for all process units are incorporated by means of friction factor correlations. For the mixing section of the ejector, a momentum balance is adopted. The lumped combustor model is based upon total conversion of all combustible compounds. Good numerical stability was obtained when using this model in contrast to the combustor model with chemical equilibrium. The assumption of total conversion is valid provided that excess oxygen is available. From a practical point of view, excess oxygen should always be available to achieve complete combustion and to avoid the formation of chemical species such as carbon monoxide. Performance maps have been incorporated to represent the behaviour of GT and fuel compressor under offdesign conditions. The GT expander model is based on the Stodola equation [6]. A dynamic shaft model is applied for the fuel compressor accounting for acceleration/deceleration of the shaft through moment of inertia of moving parts. The shaft model connecting the single-shaft GT and generator is assumed to be at steady-state because of the direct connection to the grid. The thermal power plant efficiency is obtained by

Konrad Eichhorn Colombo, O. Olav Bolland/ EnergyProcedia Procedia 1 00(2009) (2008) 431–438 000–000 K.E. Colombo, Bolland / Energy

Kth

100

3

Pg  Wc , fuel

433

(1.1)

n fuel LHV fuel

Where Pg is the generator power output, Wc , fuel specifies the work for the fuel compressor, n fuel is the mass flow of fuel, and LHV fuel indicates the lower heating value of the fuel. The whole GT power plant is implemented in the modelling software

gPROMS [7], an equation-oriented modeling tool. The whole GT power plant model consists of 13,702 equations. The system was solved by means of the standard solver DASOLV. Physical properties are based on Soave-Redlich-Kwong equation of state. A detailed model description can be found in [4, 8]. Relaxing solver settings leads to a total CPU time of less than 144 seconds for all simulations. 2.2. Simple Gas Turbine Power Plant As a reference, a simple GT power plant was modelled, which is shown in Fig.2. The simple GT power plant is based upon the same conditions with respect to air mass flow, TIT, and environmental conditions as the OMCM-based GT power plant. It should be noted that the GT matching point between the simple GT power plant and the OMCM-based GT power plant are different as a result of different mass flow ratios of air and fuel. This can be seen on Fig.1 and Fig.2.

T=288.15 Ŷ p=30 Ŷ m=1.35 xCH4=1 Ŷ Natural Gas Supply

Fuel valve

T=685.76 p=18.04 T=288.15 Ŷ Air p=1.01325 Ŷ m=63.47 SM=14.81 xN2=0.79 Ŷ Șis=83.7 xO2=0.21 Ŷ

m mass flow [kg/s] p pressure [bar] P power [MW] SM surge margin [-] T temperature [K] x mole fraction [-] Ș efficiency [-] Ŷ constant value

T=1,531 p=18.23 xN2=0.761 xO2=0.128 xH2O=0.074 xCO2=0.037

T=879.4 p=1.01325 Ŷ Exhaust P=24.35 Șmech=98.5 Ŷ

Combustor Șth,power plant=35.82

Generator

Shaft

G Compressor

Turbine

Fig. 2: Simple gas turbine power plant with values for key parameters in design. 3. Constraints Incorporating an OMCM reactor for air separation into a GT power plant increases process complexity and adds several additional operation and material constraints which need to be considered. In this respect, the OMCM itself contributes the most. Higher and lower temperature limitations must be maintained so that degradation, caused by thermo-mechanical-chemical stresses, is avoided. The HXs in the OMCM reactor are assumed to be fabricated of similar ceramic materials as the OMCM [5, 9]. Consequently, for these reactor units temperature limitations have to be met, too. Large temperature gradients in axial directions should be omitted for all monoliths to minimize the resulting stresses [10]. This constraining factor depends not only on individual material properties of the OMCM and HXs (such as thermal expansion coefficient), but on the interconnection between OMCM, HXs, sealing, and further joints. Utilising soft edge seals would lead to considerably reduction of induced stresses because the OMCM and HXs can then relax by deformation. On the other hand, it is questionable whether such sealing materials remain gas tight under variation of operation conditions. Large total pressure gradients between sweep gas and air should be avoided since the monolithic OMCM and HXs have solid walls of 0.33mm thickness [4]. Further, combustion of oxygen and methane in a highly-diluted atmosphere, mainly consisting of CO2 and H2O, is very challenging. It is assumed that stable combustor operation can be maintained with an excess oxygen mole fraction of 0.5 per cent. Rather high pressure drops in the mixing section of the ejector must be allowed for which leads to reduced power plant efficiency at design conditions. Marsano stated that the total pressure difference in the mixing section of the ejector should be larger than zero in order to guarantee correct ejector performance [11]. In this work, a margin of 0.01bar as lower limit has been defined for reason of transient inertia of the OMCM reactor. The TIT limit is assumed to be 1,513K [12]. Degradation of the OMCM may limit the

4

434

Konrad Eichhorn Colombo, OlavO.Bolland/ Procedia 00 (2008) 000–000 K.E. Colombo, BollandEnergy / Energy Procedia 1 (2009) 431–438

lifetime of the OMCM and other reactor units. Therefore, the OMCM-based power plant must be controlled in a careful manner so that all operation and material constraints are satisfied. Table 1: Operation constraints of the OMCM-based gas turbine power plant. Process Unit

Constraint

Effect

Value

OMCM

Min. temperature limit

Thermo-mechanical stresses

1,173K [9]

OMCM

Max. temperature limit

Thermo-mechanical stresses

1,323K [9]

OMCM and HXs

Max. transient heating rate

Thermo-mechanical stresses

3K/min [9]

OMCM and HXs

Max. transient cooling rate

Thermo-mechanical stresses

5K/min [9]

OMCM and HXs

Total pressure gradient

Mechanical stresses

< 1bar

OMCM and HXs

Temperature gradient per length

Thermo-mechanical stresses

10 K/cm

LTHX

Min. temperature limit

Thermo-mechanical stresses

673K [9]

HTHX

Max. temperature limit

Thermo-mechanical stresses

1,573K [9]

GT compressor

VIGV angle limit

Mass flow reduction of air

30% [13]

GT and fuel compressor

Surge limit

Stable operation

> 10%

Combustor

Excess oxygen mole fraction

Stable operation

> 0.5%

Ejector

Pressure difference in mixing section

Stable operation

> 0.01

GT expander

TIT

Thermo-mechanical stresses

1,531K [12]

4. Design Performance Key results for the GT power plant in design are shown in Fig. 1. The geometry of the ejector, monolithic OMCM and HXs in the OMCM reactor has been chosen to meet all process constraints shown in Table 1. The monolithic OMCM and HXs have a rather small operation window with respect to temperature limitations as well as total pressure gradients between the two fluids. Besides, the OMCM must be operated at process conditions where carbonate formation as well as oxidation do not occur [9]. Both mechanisms lead to irreversible membrane degradation. Carbonate formation has been incorporated as a check variable based on the stability diagram presented by [4]. At the sweep gas inlet of the OMCM, carbonate formation is very likely to occur due to the low oxygen partial pressure and high partial pressure of carbon dioxide in spite of high temperatures. One solution might be to coat critical parts of the OMCM by a protection layer [10]. However, it is expected that this protection layer would represent a resistance for the oxygen permeation. As a result less oxygen is transported through the OMCM which leads to a shift of other critical variables in the GT power plant such as excess oxygen for combustion. The OMCM and the LTHX have very low tolerance towards sulphuric compounds [9]. Hence, all these contaminants need to be removed from the natural gas. In this work, natural gas is assumed as pure methane. The total pressure in the recycle loop is calculated by means of the ideal gas law incorporating the geometric volume of individual reactor units. Further assumptions with respect to geometry of reactor units in the recycle loop can be found elsewhere [4]. 5. Part-Load Performance In times of decreased power demand the power plant moves from design to off-design conditions. In the OMCM-based GT power plant all operation and material constraints, shown in Table 1, must still be satisfied under off-design conditions. For load reduction two strategies are analysed: (i) reduction of mass flow of air in conjunction with reduced fuel supply to the afterburner while keeping the turbine exit temperature (TET) constant (TET control strategy), and (ii) reduction of fuel supply to the afterburner at constant air supply to the GT while the TET is allowed to vary (turbine inlet temperature (TIT) control strategy). The OMCM reactor should be operated at its maximum temperature to meet all operation and material constraints. Load reduction can be accomplished until one of the constraints cannot be longer maintained. Table 2 shows the set of manipulated and controlled variables of the OMCM-based GT power plant. During operation the combustion temperature cannot directly be measured but is based on measurement of the mole fraction of excess oxygen after combustion. The TIT can be derived from measurement of the TET. The GT is assumed to be a single-shaft engine with a direct connection to the grid. Load reduction by means of variable shaft speed can therefore not be applied. Instead, variable inlet guide vanes (VIGV) in front of the GT compressor manipulate the mass flow of air at a constant rotational speed following the TET control strategy. In the TIT control strategy, load reduction is accomplished by decreasing the fuel supply to the afterburner. The total pressure in the recycle loop is controlled by keeping the total pressure gradient across the BHX at a minimum. It is important to emphasise that all monoliths in the OMCM reactor are limited to total pressure gradients. Controlling the difference between the air inlet and sweep gas outlet in the BHX results in larger total pressure gradients in other monoliths because of pressure drops.

Konrad Eichhorn Colombo, O. Olav Bolland/ EnergyProcedia Procedia 1 00(2009) (2008) 431–438 000–000 K.E. Colombo, Bolland / Energy

5

435

Table 2: Set of controlled and manipulated variables of the OMCM-based gas turbine power plant. Controlled Variables

Manipulated Variables

Power output

VIGV angle in front of GT compressor

Total pressure in recycle loop

Valve opening of the pressure control valve

Turbine exit temperature

Valve opening of the fuel valve

Excess oxygen after combustion

Rotational speed of the fuel compressor

Mass of steam in sweep gas

Valve opening of the steam valve

Table 3 shows key results of the OMCM-based GT power plant in design and maximum load reduction for the TIT and TET control strategy in comparison to the simple GT power plant. The operation conditions with respect to TIT, mass flow of air as well as environmental conditions are equal for both GT power plants. In the simple GT power plant, following the TET control strategy, the load can theoretically be reduced until the mass flow of air entering the GT compressor is 30 per cent of the design value. However, for reason of comparison simulation with a load reduction equal to that of the OMCM-based GT power plant were performed. During load reduction of the OMCM-based GT power plant the fuel and steam supply to the OMCM reactor is controlled to maintain a combustion temperature of 1473K. The total pressure in the recycle loop is controlled at the sweep gas outlet of the bleed-gas HX (BHX) to hold the pressure difference between air and sweep gas at zero. The limiting factor for the TET control strategy is the total pressure difference in the mixing section of the ejector, whereas for the TIT control strategy the excess oxygen mole fraction after combustion is the restricting variable. The TET control strategy leads to increased load reduction when compared to the TIT control strategy. Further, the excess mole fraction of oxygen after combustion is increased, leading to better performance of the combustor. Higher excess oxygen rates in the sweep gas, however, require additional purification in the CO2 separation and compression process because of increased oxygen. The surge margin (SM) of the GT compressor is reduced but is sufficiently high to avoid unstable operation. Maintaining a high TET at part load gives higher work output and increased total efficiency when the exhaust gas from the GT is utilized in a steam cycle. A drawback of the TET control strategy is the increased total pressure difference across the OMCM. Specific CO2 emissions are higher for the TET control strategy because of higher fuel supply to the afterburner. In conclusion, the TET control strategy is generally superior to the TIT control strategy. Load reduction by lowering the maximum temperature in the OMCM reactor has been rejected as a possible solution since material constraints got worsen. Ultimately, this leads to more rapid damage of OMCM reactor units. It should be noted that the production of steam, which is required for the ejector as an additional actuating fluid in the OMCM reactor, is not taken into account for the calculation of the thermal efficiency. The thermal efficiency is therefore lower as shown in Table 3.

Table 3: Simulation results for key variables of the simple and OMCM-based GT power plant for the TIT and TET control strategy.

6

436

Konrad Eichhorn Colombo, OlavO.Bolland/ Procedia 00 (2008) 000–000 K.E. Colombo, BollandEnergy / Energy Procedia 1 (2009) 431–438

Variable

Full load

Minimum load TIT control strategy

Minimum load TET control strategy

GT power plant

simple

OMCM-based

simple

OMCM-based

simple

Thermal efficiency [%]

35.82

32.9

34.81

31.5

33.79

OMCM-based 31.1

Power output [MW]

24.35

19.33

20.22

16.05

17.14

13.6

total pressure drop in ejector mixing section [bar]

-

3

-

2.62

-

0.12

Average total pressure diff. across OMCM [bar]2

-

-0.25

-

-0.42

-

-0.81

14.81

14.6

20.25

18.8

10.32

10.7 45.93

SM of GT compressor [%] SM of fuel compressor [%]

-

28.9

-

35.3

-

100

100

100

100

77.4

78.5

Pressure ratio of GT compressor [-]

17.81

17.85

17

17.2

13.43

13.62

Specific CO2 emission per MW [kg CO2/MW]

0.153

0.036

0.158

0.016

0.162

0.0175

Turbine inlet temperature [K]

1531

1531

1412.5

1426.2

1458.6

1460.4

Turbine exit temperature [K]

879.4

874.25

809.75

813.1

879.4

874.25

-

1.04

-

0.55

-

1.17

VIGV opening [%]

Excess oxygen after combustion [%]

6. Transient Performance 6.1. Time Scale of Process Components Time scales of the monolithic OMCM and HXs, respectively have been analysed previously [8]. The time scale of heat transfer between the two fluids in the monoliths is in the order of minutes. The OMCM reactor and pipes are assumed to be coated with a high-performance insulation material with low thermal conductivity. Consequently, heat propagation to the environment is very slow and takes several hours. With such insulation, heat losses can be reduced. In addition, shorter start-up times after a short stop are required for the power plant since the OMCM reactor can be maintained at fairly high temperature. Griffin, et al. [15] state that the required time for complete methane combustion in a highly diluted atmosphere is in the range of milliseconds. However, this time range (<1s) has not been considered in the power plant model in order to avoid the problem of numerical stiffness. Besides, pressure and mass propagation in fluid machinery (GT and ejector) are assumed to occur instantaneously. Dynamics of the turbomachinery are not incorporated in the shaft model of the GT because of the direct connection to the grid. But the shaft model connected to the small fuel compressor and driving motor is based on a transient power balance. Table 2 shows the time scale of process components in the GT power plant. Table 2: Time scale of process components in the gas turbine power plant.

2

Response in Process Components

Time Scale [s]

Oxygen permeation in OMCM3

Instantaneous [8]

Pressure and mass propagation in fluid machinery (GT, ejector)

Instantaneous

Thermal in combustion

Milliseconds [15]

Mass propagation in monolithic OMCM and HXs

Seconds [8]

Shaft inertia of fuel compressor

Seconds

Thermal in monolithic OMCM and HXs

Minutes [8]

Thermal in insulation material (OMCM reactor, pipes)

Hours [8]

The total pressure difference is defined as ǻpOMCM = pair - psweep gas. Start-up times between 5 to 10 hours have been reported, which depend on material and thickness of the membrane as well as oxygen permeation control by surface reactions. However, once the steady-state oxygen permeation is established at high temperatures, it remains stable during relatively fast changes of operation conditions [16]. 3

Konrad Eichhorn Colombo, O. Olav Bolland/ EnergyProcedia Procedia 1 00(2009) (2008) 431–438 000–000 K.E. Colombo, Bolland / Energy

7

437

6.2. Transient Load Reduction

110

40

110

40

38

100

38

90

36

34

load [%]

36 90

efficiency [%]

load [%]

100

80

34

load efficiency

80

70 0

50

100

150

200

efficiency [%]

The GT in a 250MW combined cycle can be shut-down in less than 23 minutes [17]. Consequently, less than 2.3 minutes are required for a load reduction of 10 per cent by the GT. Fig.3 shows the transient load reduction of the simple GT power plant for the TIT and TET control strategy. It should be mentioned that the simple GT power plant is capable of considerably farther reduction of load. However, for reason of comparison equal ratios of load have been assumed for both power plants.

load efficiency

32

70

30

60

250

32

30 0

100

200

time [s]

300

400

time [s]

Figure 3: Transient load reduction and thermal efficiency of the simple GT power plant for TIT control strategy (left) and TET control strategy (right). The OMCM-based GT power plant exhibits rather slow dynamics during load reduction as already indicated by Hamrin [18]. A load reduction of 10 per cent takes more than 32 minutes for the TIT control strategy and more than 64 minutes following the TET control strategy. This slow transient behaviour can be explained by the large amount of mass which is being recirculated. Faster load reduction would be possible from a stress-depending point of view since cooling rate limitations, which are given in Table 1, are not exceeded. The limiting factor for stable operation is overshooting of internal variables such as excess oxygen in the combustor. The transient load reduction for the OMCM-based GT power plant, following the TIT and TET control strategy, is shown in Fig. 4.

120

120

34

110

34

110 32

28

70 26 60 1e+0

1e+1

1e+2

1e+3

time [s]

1e+4

1e+5

30

90

80

load efficiency

efficiency [%]

load efficiency

80

100

load [%]

30

90

32

efficiency [%]

load [%]

100

28

70 26 60 1e+0

1e+1

1e+2

1e+3

1e+4

1e+5

time [s]

Figure 4: Transient load reduction and thermal efficiency of the OMCM-based GT power plant for TIT control strategy (left) and TET control strategy (right). 7. Conclusion The OMCM-based GT power plant features a high degree of complexity and adds several operation and material constraints when compared to a simple GT power plant. Hence, part load operation of the OMCM-based GT power plant is a challenge. Further, design performance indicates an efficiency drop of 3 percentage points. Load reduction for both power plants is presented by means of two different control strategies. In the first control strategy the TET is kept constant while reducing the air supply to the GT compressor by means of VIGVs. The second control strategy uses reduced fuel supply to the afterburner whereas the TET is allowed to float. The mass flow of air to the GT is kept constant. For the OMCM-based GT power plant, the

8

438

Konrad Eichhorn Colombo, OlavO.Bolland/ Procedia 00 (2008) 000–000 K.E. Colombo, BollandEnergy / Energy Procedia 1 (2009) 431–438

TET control strategy is superior because of the capability to reduce the load farther as well as improved performance in terms of operation and material constraints, in particular of the excess oxygen mole fraction. Transient simulation reveals the slowness of the OMCM reactor which is a result of the large amount of recirculated gas. While load reduction of 10 per cent in the simple GT power plant can be undertaken in less than 2.3 minutes, the OMCM-based GT power plant requires more than 64 (TET control strategy) and 32 minutes (TIT control strategy), respectively. In addition, a maximum load reduction of approximately 30 per cent can be realized in the OMCM-based GT power plant due to its operation constraints. It should also be noted that the supply of steam in a well-controlled manner is required for the OMCM-based GT power plant. A suitable control strategy is necessary for load control where all constraints are sustained. That is subject of current investigation. In conclusion, the OMCM-based GT power plant has the potential of reducing CO2 emissions in power generation. However, operation of this power plant is very challenging due to its additional constraints when compared to mature GT power plants. 8. Acknowledgement This publication forms a part of the BIGCO2 project, performed under the strategic Norwegian research program Climit. The authors acknowledge the partners: StatoilHydro, GE Global Research, Statkraft, Aker Clean Carbon, Shell, TOTAL, ConocoPhillips, ALSTOM, the Research Council of Norway (178004/I30 and 176059/I30) and Gassnova (182070) for their support. The first author acknowledges the discussions with his colleague Lars Olof Nord on gas turbine performance and thanks for his valuable comments on this paper. 9. References [1] Griffin, T., Sundkvist, S.G., Åsen, K., Bruun, T.; Advanced zero emission gas turbine power plant; Journal of Engineering for Gas Turbines and Power, v 127, n 1, January, 2005, p 81-85. [2] Kvamsdal, H.M., Jordal, K., Bolland, O.; A quantitative comparison of gas turbine cycles with CO2 capture; Energy, v 32, n 1, January, 2007, p 10-24. [3] Gellings, P.J., Bouwmeester, H.J.M.; The CRC handbook of solid state electrochemistry; CRC press; 1997. [4] Eichhorn Colombo 2, K., Bolland, O., Stiller, C.; Design and Part-Load Performance of an Oxygen Mixed Conducting Membrane-based Combined Power Cycle with an Emphasis on Operation Constraints; to be published. [5] Julsrud, S. and Viegeland, B.E.; Ceramic heat exchangers; United States Patent Application 20050009686; 2005. [6] Traupel, W.; Thermische Turbomaschinen, Zweiter Band, Geänderte Betriebsbedingungen, Regelung, Mechanische Probleme, Temperaturprobleme; third edition, Springer-Verlag; 1982. [7] General Process Modelling and Simulation Tool; v.3.1.3, Process Systems Enterprise Ltd., London, www.psenterprise.com; 2008. [8] Eichhorn Colombo 1, K., Imsland, L., Bolland, O., Hovland, S.; Dynamic Modelling of an Oxygen Mixed Conducting Membrane and Model Reduction for Control; submitted to Journal of Membrane Science. [9] Sundkvist, S.G., Julsrud, S., Vigeland, B., Naas, T., Budd, M., Leistner, H. and Winkler, D.; Development and testing of AZEP reactor components; International Journal of Greenhouse Gas Control, v 1, n 2, April, 2007, p 180-187. [10] Hendriksen, P.V., Larsen, P.H., Mogensen, M., Poulsen, F.W., Wiik, K.; Prospects and problems of dense oxygen permeable membranes; Catalysis Today, v 56, n 1-3, Feb 25, 2000, p 283-295. [11] Marsano, F., Magistri, L., Massardo, A.F.; Ejector performance influence on a solid oxide fuel cell anodic recirculation system; Journal of Power Sources, v 129, n 2, Apr 22, 2004, p 216-228. [12] Rodrigues, M., Walter, A., Faaij, A.; Performance evaluation of atmospheric biomass integrated gasifier combined cycle systems under different strategies for the use of low calorific gases; Energy Conversion and Management, v 48, n 4, April, 2007, p 1289-1301. [13] Kim, T.S. and Hwang, S.H.; Part load performance analysis of recuperated gas turbines considering engine configuration and operation strategy; Energy, v 31, n 2-3, February/March, 2006, p 260-277. [14] Caro, J.; Membranreaktoren für die katalytsiche Oxidation; Chemie Ingenieur Technik; Volume 78 Issue 7, Pages 899 – 912; 2006. [15] Griffin, T., Winkler, D., Wolf, M., Appel, C., Mantzaras, J.; Staged catalytic combustion method for the advanced zero emission gas turbine power plant; Proceedings of the ASME Turbo Expo 2004, v 1, Proceedings of the ASME Turbo Expo 2004. Volume 1: Combustion and Fuels; Education, 2004, p 705711. [16] Zhang, W., Smit, J. van Sint Annaland, M., Kuipers, J. A.M.; Feasibility study of a novel membrane reactor for syngas production: Part 1: Experimental study of O2 permeation through perovskite membranes under reducing and non-reducing atmospheres; Journal of Membrane Science, Volume 291, Issues 12, 15 March 2007, Pages 19-32. [17] Kehlhofer, R., Bachmann, R., Nielsen, H., Warner, J.; Combined-cycle gas & steam turbine power plants; Penn Well Publishing Company; 1999. [18] Hamrin, S.; Control of a gas turbine with hot-air reactor; United States Patent Application 20060230762; 2004.