- Email: [email protected]
Model based scale-up study of the calcium looping process
Fuel 115 (2014) 329–337
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Model based scale-up study of the calcium looping process Jaakko Ylätalo ⇑, Jouni Ritvanen, Tero Tynjälä, Timo Hyppänen Lappeenranta University of Technology, LUT Energy, P.O. Box 20, 53851 Lappeenranta, Finland
h i g h l i g h t s Post-combustion calcium looping unit was retrofitted for a 250 MWth combustor. Dimensioning of the unit was done with a 1-D dynamic model. Five flue gas load cases were analysed from 100% to 0% load. Flue gas and solid circulation played a key role in control of the system. Heat transfer and solid transfer design are important in large scale units.
a r t i c l e
i n f o
Article history: Received 17 June 2013 Received in revised form 10 July 2013 Accepted 11 July 2013 Available online 25 July 2013 Keywords: Calcium looping process 1-D modeling CO2 capture Fluidized bed reactor Scale-up
a b s t r a c t A 1-D dynamic calcium looping model was applied to a large scale calcium looping concept capturing CO2 from a 250 MWt power plant. Several new features were added to the existing model frame in order to successfully simulate the large scale unit. Models were needed for new material fractions, such as ash and CaSO4, sulfur capture and heat transfer in the solid return system. The plant was dimensioned based on the experience from large CFB units and the heat transfer design was evaluated based on initial simulations of the design case. The unit was then simulated in five load scenarios ranging from full load to zero load, no flue gas flow to the carbonator. The scale-up of the process is feasible from the model’s point of view keeping in mind the assumptions and simplifications made in the modeling. The results from the simulations confirmed that successful operation of a large scale calcium looping unit requires good heat transfer design including cooling of the hot solids coming the calciner. Also the recirculation of flue gas in both reactors is necessary to ensure the sufficient fluidization for different flue gas flows from the source combustor. Solid circulation control is also critical because it affects heavily the thermal balance of the system as well as the capture efficiency and the CO2 balance. Results show that utilizing the experience gained from the large CFB-units coupled with new innovations, the scale-up of this process could be feasible in the near future. The ability to operate flexibly in different modes could give this technology an advantage needed for wide industrial utilization of the process. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The scientific community around the world is almost unanimous that the actions of mankind have accelerated the release of greenhouse gases into the world’s atmosphere and consequently increased the chance of rapid climate change in the coming decades. The general opinion among legislators and the scientific community is that swift actions should be taken to prevent global warming by cutting down greenhouse gas emissions [1]. It is apparent that this cannot be achieved by single actions but by applying a variety of methods ranging from clean energy production to population and economy growth control. Energy production is one of the largest contributors to greenhouse gas emissions but
⇑ Corresponding author. Tel.: +358 40 171 4003. E-mail address: jaakko.ylatalo@lut.fi (J. Ylätalo). 0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2013.07.036
it is also approachable from a technical perspective so the potential to cut down emissions quickly is high. The demand for energy is increasing year by year and actions should be taken to reduce the emissions from fossil fuels. This can be done by utilizing a variety of methods simultaneously, carbon capture and storage as one of the most promising methods to cut down emissions quickly from existing stationary sources. Several carbon capture methods are being developed to answer the need to reduce stationary emissions from fossil fuel power plants. Oxy-combustion and amine-based post-combustion capture methods are technologically more mature and in the demonstration phase. Second generation capture technologies like chemical looping and calcium looping are also developing quickly and are advancing in the pilot stage [2–4]. Calcium looping is a technology that uses lime to absorb CO2 from flue gases in a fluidized bed reactor and release it in a second fluidized bed regenerator to produce a clean CO2 flow
330
J. Ylätalo et al. / Fuel 115 (2014) 329–337
to the storage [5,6]. Post-combustion calcium looping and all of its variations [7–9] are promising technologies for reducing CO2 emissions from coal-fired power plants because the availability of limestone and potential for high CO2 and SO2 capture rate. Also the technological advancements in calcium looping, like thermal coupling of the combustor and carbonator or high performance sorbents, can increase process efficiency and greatly reduce the energy penalty due to carbon capture. [10,11]. One of the major issues in the second generation capture technologies is the scale-up to industrial scale units. The construction of these industrial units requires a large capital investment. However, for the calcium looping technology to succeed, operational experience from large scale units is required, before new technologies can be utilized in commercial scale. This can be partly achieved by using a variety of computational tools available for process modeling and simulation. Utilizing a variety of tools, different phenomena can be researched, for example, CFD calculations enable the study of gas–solid interactions and dual fluidized bed behavior but are computationally very strenuous [12]. Process modeling tools like AspenÒ or IPSEproÒ help to understand the whole process and how different phenomena affect each other and are computationally more forgiving but rely on several assumptions [13–15]. Both approaches are required to comprehensively understand the process and gradually transfer systems to the industrial scale. In this paper, a modeling approach is used to scale-up the postcombustion calcium looping process to capture CO2 from a conceptual 250 MWt power plant. A 1-D dynamic process model is used to dimension the reactors and simulate the interconnected system ensuring that the operational parameters are feasible and to study the behavior of the unit in different load scenarios. Several studies have been published [16–18] where large scale calcium looping systems have been simulated but none of the studies has incorporated reactor design, thermal integration and overall loop simulation in one study. 2. Modeling approach The calcium looping process is based on the partially reversible reaction between calcium oxide and carbon dioxide. CO2 from the flue gases of a combustor are absorbed in a fluidized bed of lime, carbonator. Entrained solids containing CaCO3 are transferred to a regenerator, calciner, where CO2 is released from the solid material. Regenerated CaO is entrained back to the carbonator where the loop is closed. A 1-D dynamic calcium looping model will be applied to this looping process [19]. The 1-D model will be used to study the dimensioning and operation of the reactor units to determine whether there are any technical or engineering problems in the scale-up. 2.1. The 1-D calcium looping model The 1-D dynamic calcium looping model frame has been built piece by piece adding important phenomena and coupling reactors and loop components. The model frame is solved in the MatlabÒ Simulink environment using built-in ordinary differential equation solvers. The model is capable of both dynamic and steady-state simulations. The carbonator and calciner are spatially discretized into 1-D control volumes in which mass and energy balances are calculated each time step using the first order upwind scheme. The reactors are interconnected in the Simulink environment and 0-D models of the loop seals, cyclones and standpipes exist in the model frame. The reactor model frames incorporate several fluidized bed phenomena like solid behavior in a circulating fluidized bed, solid suspension heat transfer, core-annulus calculation, solid
fuel combustion, sulfur capture, ash accumulation and cyclic carbonation and calcination. For solid circulation and suspension, semi-empirical models have been applied from the literature. The vertical division of solids in the reactor is modeled with a modified approach presented by Johnsson and Leckner [20]. The equation was modified by leaving out the static bed term. Solid flow out of the reactor is modeled by a semi-empirical equation including exit density of the reactor, slip velocity of solids in the exit and exit channel dimensions [19]. Both the solid density profile and solid flow out of the reactor are tied to the total solid inventory of the reactor, particle properties and gas velocity profile of the reactor. The rates of chemical reactions have been defined from the best available knowledge from the literature. The reaction rate for carbonation was applied from the work of Shimitzu et al. [5]. Carbonation is limited by the aging of the lime by defining maximum carbonate content for the solids. Calcination reaction rate was acquired from the model proposed by Fang et al. [21]. The combustion model has been built based on an approach presented by Myöhänen and Hyppänen [22]. The combustion model divides the input fuel into char, volatiles, ash and moisture. Fuel moisture is released immediately to the gas balance. A time constant can be defined for the release of volatiles. Volatiles are combusted in the gas balance based reaction rates acquired from the literature. Char and ash are divided into char balance and solid balance. Char combustion is modeled with a simple model homogeneous combustion model. Ash is considered inert and will be a part of the system solids with homogeneous properties. The core-annulus model enables the transfer of material and heat from the top part of the reactor to the bottom part but the material in the annulus is not considered reacting. Mass and energy balances for the core and annulus are solved separately. Solids flow in the core upwards and downwards in the annulus. The coreannulus model can be used to emulate the internal flow of solids in a fluidized bed reactor. The amount of solid flow in the annulus is determined based on experimental results keeping in mind that the solid volume fraction cannot exceed 0.6 and the solid velocity in the core cannot exceed gas velocity. In addition to the coreannulus approach, mixing of energy in consecutive control volumes can be modeled with a turbulent diffusion. The energy balance incorporates several phenomena present in a fluidized bed reactor. In addition to energy transported alongside solid and gas flows, heat losses through the walls can be modeled as well as heat transfer to membrane walls with or without refractory. Single heat sinks or sources can be defined in each control volume. Heat produced or consumed by chemical reactions is also included in each control volume energy balance. The model has been validated both with lab-scale equipment as well as with 1.7 MW pilot scale experimental results [2,23]. Alongside that, validation and comparison between the 1-D and a 3-D steady-state calciner model has been done [24]. Generally, the maturity of the model starts to near a tool suitable for scale-up purposes. The 1-D model tool can be used to evaluate the solid circulation between the reactors, the amount of heat transfer surface required inside the reactors, performance in different loads, the amount of make-up required for successful CO2 and SO2 capture without performance loss and many other features. Studying these features will point out whether the calcium loop is technically feasible for the calculated case. 2.2. Improvements and updates to the model frame The model frame has been improved since the last publications [19,24] by adding several features to more accurately describe the process. The calculation of material fractions CaSO4 and ash has been added to the solid balance in order to model sulfation and ash accumulation. The equations to calculate the solid material
J. Ylätalo et al. / Fuel 115 (2014) 329–337
fractions remain the same as in Ylätalo et al. [19], only now instead of two materials, CaO and CaCO3, there are four. The sulfation rate equation has been implemented from De-Souza Santos [25] taking into account only indirect sulfation, the reaction between calcium oxide and sulfur dioxide. This assumption is made because the calciner has an abundance of CaO for sulfur capture compared to CaCO3 even though the calciner can operate near the equilibrium. 8810 T
r sulf ;CaO ¼ 4:9 103 ð3:843T þ 5640Þe
X CaO xSO2 xO2
ð1Þ
where rsulf,CaO is the sulfation rate of calcium oxide (1/s), T is the surrounding temperature (°C), XCaO is the calcium oxide molar fraction (-), xSO2 and xO2 are the molar fractions of SO2 and O2 in the surrounding gas. Tracking material fractions can also be used to design solid material purge and make-up flows for the system. Another feature is the ability to extract heat from the solid return system by using simple heat balance calculations in the loop seal, stand pipe and cyclone computational blocks. This addition enables better thermal design of the system which will be demonstrated in the following sections. The solved equation for the solid return system is
331
Table 1 Basic information of the 250 MW simulation case and some essential simulations parameters. Combustor thermal power (MW) Flue gas mass flow (kg/s) Flue gas CO2 (wt%) Flue gas SO2 (wt%) Flue gas temperature (°C) Designed capture efficiency (%) Calciner fuel flow (kg/s) Fuel LHV (MJ/kg) Oxidant flow (kg/s) Make-up flow (kg/s) Estimated maximum average CaCO3 fraction (wt%) Fuel char (wt%) Fuel volatiles (wt%) Fuel moisture (wt%) Fuel ash (wt%) Average particle size (lm) Average particle density (kg/m3) Turbulent dispersion coefficient (-) Circulation coefficient (-) Wall layer flow speed (m/s)
250 115.92 21.16 0.17 120 80 9.229 30 26 11.15 41 67.6 22.5 7.5 2.4 200 1800 0.3 0.3 0.01
Q removed ¼ qm;s cp;s ðT in;s T element Þ ¼ hin Awall ðT element T wall Þ ¼
kwall Awall ðT wall T out Þ swall
ð2Þ
The thermal power removed from the system is represented by Qremoved (W), qm,s is the solid flow through the system (kg/s), cp,s is the solid heat capacity (J/kgK), Tin,s is the solid temperature entering (°C) and Tblock is the internal temperature of the 0-D block (°C) referring for example to the loop seal internal temperature. The heat transfer coefficient of the block is hin, approximately 100– 200 W/m2K, Awall is the inside wall area or the block and Twall is the wall temperature. The heat conductivity of the wall material is kwall (W/mK) and swall (m) is the thickness of the material and Tout is the outer temperature of the block (°C) for example heat transfer surface or outside metal temperature. The loop seal, standpipe and cyclone can be insulated with refractory or they can be equipped with heat transfer surfaces. A PID-controller has been constructed to control the calciner purge to handle the total solid inventory of the system more efficiently and to limit the accumulation of ash and CaSO4. PID-controller follows the calciner inventory and changes the purge according to the desired calciner inventory defined in the input. Make-up of the system is kept constant. The controller slows down the simulation significantly but increases stability which is an acceptable trade-off. The PID-controller resembles the actions of an operator who would follow the pressure difference of the bed and adjust the calciner purge. Further analysis of the system control is not done in this study: the controller is used to find a stable solution faster than by iterating the purge by hand, although the control loop increases calculation times. 3. Results 3.1. Introduction of the scale-up case A conceptual power plant with a thermal power of 250 MW was selected as the scale-up case. Several factors affected the size of the selected case. The pre-industrial size poses less financial risk but can be successfully extrapolated to a larger scale. Also the retrofitting with the calcium loop doubles the thermal output of the plant. General operation parameters of the power plant and the calcium loop are listed in Table 1 as well as the composition of the primary fuel and some essential simulation parameters. Currently the model can handle only one fuel so high quality coal is selected for the combustor and the calcium loop although power plants tend to use several fuels in combination. The calcium looping unit was
modeled as a retrofit unit without steam cycle or thermal integration to the 250 MW boiler.
3.2. Dimensioning and thermal design of the reactors The flue gas flow rate of the carbonator can be used to dimension the carbonator reactor once the average particle size is determined for the system. For the selected fuel and air-ratio of 1.2, the flue gas flow is 116 kg/s. Studies have shown that calcium looping systems tend to have a small particle size (<100 lm) due to fragmentation in the initial calcination [26]. With this particle size, it is quite difficult to operate in the conventional CFB-region. That is why it is more practical to use a higher average particle size commonly found in CFB-units. If the average particle size is fixed to 200 lm, the maximum superficial gas velocity can be set to 6 m/s. Both particle sizes are in the Geldart A class but the 100 lm particles approach the pneumatics transport region. A larger particle size could be achieved by using more durable limestone against fragmentation or increasing initial feed size. Calculating from the carbonator gas mass flow, approximate density and the maximum velocity, the cross section of the carbonator will be around 44 m2. Because the gas velocity in the carbonator is decreasing due to the adsorption of CO2 by the CaO, the use of sloped bottom section, commonly found in CFB boilers, is not appropriate. Also using secondary flue gas inputs is not necessary because that will only cause some of the flue gas CO2 to bypass the dense, reactive part of the carbonator bed. The rectangle like cross section commonly used in CFB-boilers is not necessary but in order to keep structural integrity with the calciner, the carbonator width is determined to be the same as the calciner. The height of both reactors could be defined to maximize the residence time of gas and solids but in practice, the height the cyclone, standpipe and loop seal combination sets the maximum height of the system. In this case, a value of 35 m could be close to the reality when examining existing CFB boilers. In the thermal design it was determined that an external heat exchanger would be the best solution to remove heat from the incoming solids from the calciner and carbonation reaction. An evaporator surface in the carbonator would be inflexible in terms of cooling in lower loads. Initial simulations also pointed out that, if the carbonator bed is refractory protected due to the very high erosion rate of the dense bed, the heat transfer rate in the bed is not enough to cool the solids to 650 °C. Using in-bed heat transfer tubes would quickly cool the solids but it would also lead to fast
332
J. Ylätalo et al. / Fuel 115 (2014) 329–337
Fig. 1. Concept image of the large calcium looping unit with dimensions and material flows.
erosion of the tubes. The most reactive bed part would be outside of the equilibrium which would cause trouble in the capture. It was determined that the cooling the solids before feeding them into the carbonator would ease the heat transfer load in the carbonator bottom part. In addition, the equipment feeding the solids are also under great thermal stress and cooling of them is necessary. Location of external heat exchanger in the solid train feeding the carbonator is presented in Fig. 1. In the model the external heat exchanger is executed by cooling the incoming solids to a low temperature (550–600 °C). In an actual unit, the external heat exchanger could receive solids from the carbonator to cool the reactor more efficiently. Calciner dimensions can be evaluated similarly keeping in mind that the calciner gas flow consists of the oxygen required for combustion, CO2 from the recirculation gas and the CO2 released from the calcination of CaCO3 in the circulated material and make-up flow. A superficial velocity of 6 m/s in the freeboard was selected as a boundary condition and the fuel flow was defined from a simple 0-D balance, with these assumptions the gas flow out of the calciner is around 108 kg/s. When the operation temperature is around 950 °C, the freeboard cross section is 60 m2. The primary gas flow can be adjusted by reducing the amount of recirculation gas but it has to be kept in mind that the amount of oxygen in the gas should not rise over 40 vol%, which could lead to too high combustion temperatures. In the calciner, a sloped section is necessary because the difference between the freeboard and grid gas mass flows is significant. By evaluating the grid gas mass flow and the exiting gas flow relation and the average time taken to reach the gas flow to fully develop, it was determined that the sloped section will be 5 m high and the grid cross section will be 43 m2. With this dimensioning, the calciner velocity should remain constant along the reactor height. The width of the calciner was set to 3.8 m to keep the fuel, gas and solid penetration sufficient. Rectangular cross section commonly used in CFB-units was also selected for the calciner. This will set the length of the reactor to 15.8 m and the width of the grid section, 2.7 m, because the length is not changed in the grid. The carbonator width will be 3.8 m and depth 11.5 m and both reactors will be structurally connected at the ends, Fig. 1. This structural connection enables the construction of the solid return system alongside the reactor and the feeding of a portion of the solids back to the original reactor and transferring the remaining solids to the next reactor. The wall between the
reactors is insulated to prevent heat transfer between reactors. Also this solid train type of solution enables the division of solid feeding along the reactor length to even out the active solid concentration in the reactor. This kind of solution can be modularized: larger units could be built by adding consecutive carbonator and calciner blocks, increasing the length. The main dimensions of the reactors and the thermal design solutions are listed in Table 2. Both reactors are fitted with the ability to recirculate solids and gas. The reason for this is explained in the following sections. The calciner is fluidized with the mixture of recirculation gas and oxygen. The carbonator can be fluidized solely with flue gas from the combustor or with a combination of recirculation gas from the carbonator or just with air. 3.3. Full load results The simulation results for the full load case and design point of the plant can be seen in Figs. 2–5. Fig. 2 describes the overall balance of the system. The make-up flow for the case was determined based on the fuel ash and sulfur content, achieving an average activity of 41 wt% in the system. With the 6 m/s fluidization velocity, the solid mass flow out of the reactors is much larger than needed for moderate capture therefore 40% of the solids is recirculated in both reactors. The fluidization velocity of the calciner is purposefully lower than the carbonator velocity because it helps to control the overall solid inventory of the system. If the calciner velocity becomes higher and solids start to move to the carbonator, there is a danger that the whole reactor is drained from solids.
Table 2 Main inner dimensions of the carbonator and calciner reactors and cooling surface parameters of the large scale calcium looping plant. Dimension (inner)
Carbonator
Calciner
Height of reactor (m) Cross section of grid (m2) Cross section of freeboard (m2) Width of grid (m) Width of freeboard (m) Length (m) End of sloped section height (m) Elevation to exit channels (m) Thermal design
35 44 44 3.8 3.8 11.5 None 33 External HE
35 43 60 2.7 3.8 15.8 5 33 Insulated
333
J. Ylätalo et al. / Fuel 115 (2014) 329–337
Exit gas flow [kg/s]
95
Exit gas CO2 [w-%]
3.88
Exit gas flow [kg/s] Exit gas CO2 [w-%]
141
Exit gas O2 [w-%]
4.39
Exit gas O2 [w-%]
Exit gas N2 [w-%]
87.67
Exit gas N2 [w-%]
0.20
88.66 3.76
Exit gas H2O [w-%]
4.07
Exit gas H2O [w-%]
7.22
Exit gas SO2 [w-%]
0.00
Exit gas SO2 [w-%]
0.00
Exit gas T [°C]
654
Exit gas T [°C] Carbonator solid exit
Inventory [kg] Inventory [kg/m2]
50398 1145
Calciner solid exit
Solid mass flow out [kg/s]
362
Solid mass flow out [kg/s]
Solid flux out [kg/m2 s]
8.2
Solid flux out [kg/m2 s]
CaO [w-%]
43.7
CaO [w-%]
299 5.0
Inventory [kg]
78.2
Inventory [kg/m2]
Ecarb [%]
85
CaCO3 [w-%]
37.9
CaCO3 [w-%]
0.1
Ecalc [%]
ESO2 [%]
100
CaSO4 [w-%]
9.6
CaSO4 [w-%]
11.2
ESO2 [%]
Xave [w-%]
0.41
Ash [w-%]
8.8
Ash [w-%]
10.5
Fuel mass flow [kg/s]
Recirculation [-]
0.4
Recirculation [-]
0.4
Make-up [kg/s]
Solid flow between [kg/s]
217
Solid flow between [kg/s]
179
Purge [kg/s]
Total cooling [MW]
Flue gas flow [kg/s] Flue gas CO2 [w-%]
58
894
116
Primary gas flow [kg/s]
21.16
Primary gas CO2 [w-%]
35091 584 99 98 9.23 11.15 8.94
26 0.00
Flue gas O2 [w-%]
3.63
Primary gas O2 [w-%]
100.00
Flue gas N2 [w-%]
71.72
Primary gas N2 [w-%]
0.00
Flue gas H2O [w-%]
3.33
Primary gas H2O [w-%]
0.00
Flue gas SO2 [w-%]
0.17
Primary gas SO2 [w-%]
0.00
Flue gas T [°C]
120
Primary gas T [°C]
Recirculation [kg/s]
0.00
Recirculation [kg/s]
82.13
150
Recirculation T [°C]
120
Recirculation T [°C]
120
Fig. 2. Collected 0-D results of the full load simulation case.
35 35
30
30
25
Height [m]
25
20
20
15
15
10
10
5
5
0 0.00
0.10
0.20
Solid volume fraction [-]
0.30
0 550
600
650
700
750
Temperature [°C]
Fig. 3. 1-D plots of the carbonator solid volume fraction and temperature in the full load case.
When the purge and solid control is on the calciner side, it is beneficial to keep the velocity difference favoring the calciner, Fig. 5. Fig. 3 displays the solid volume fraction profile and temperature profile of the carbonator. Especially from the temperature profile we can see that the temperature conditions are far from ideal even in the 1-D model. The entering solids are cooler than the objective temperature of the reactor. Total thermal power removed from the external heat exchanger and solid train is 58 MW. Fig. 4 presents the corresponding solid volume fraction and temperature profiles
for the calciner. Comparing the solid volume fraction profiles, it can be seen that the carbonator has a higher inventory than the calciner which could be a result of the higher CaCO3 content in the carbonator. The CaCO3 concentration in the solids is quite close to the average activity achievable because recirculation increases the residence time of the solids in the reactor. Another interesting observation is that the calciner is running in lower temperatures than the literature suggests. However, very good calcination efficiency is achieved. This could be a result of
J. Ylätalo et al. / Fuel 115 (2014) 329–337
Height [m]
334 35
35
30
30
25
25
20
20
15
15
10
10
5
5
0 0.00
0.10
0.20
0.30
0 800
Solid volume fraction [-]
850
900
950
1000
Temperature [°C]
Fig. 4. 1-D plots of the calciner solid volume fraction and temperature in the full load case.
Calciner
Carbonator 35
30
Height [m]
25
20
15
10
5
0 3.000
4.000
5.000
6.000
7.000
Gas velocity [m/s] Fig. 5. 1-D velocity profiles of the carbonator and calciner reactors in the full load case. The calciner velocity is lower than the designed 6 m/s because the temperature is lower than originally intended especially in the lower section of the reactor.
the 1-D control volumes is quite close to the equilibrium temperature of the calcination reaction, which could cause some problems in a real unit because of the lateral temperature profiles present in a real large scale unit. In the temperature profile the insertion point of the cold solids can be seen as a drop of the temperature in the bottom region. Also the fuel insertion point close to the height of 5 m increases the local temperature by 50 K. The calcination reaction is temperature driven, which means that in the calciner the solid material has a long time to react compared to the carbonator, in which it is critical to achieve good gas–solid contact in the dense bottom region of the reactor. In other words, the calciner operation is not so sensitive to local areas of temperatures below the calcination temperature in the lower part of the reactor. The performance of the system is higher than originally intended. The capture efficiency is 85%, when the design value was 80%. The sulfur capture is also almost 100% in both reactors. Fuel flow of the calciner has been evaluated to 9.23 kg/s which is equal to 277 MW thermal power. The ratio of the calciner thermal power from total thermal power, Qcalciner/(Qcalciner + Qcombustor), is around 53% in full load.
3.4. Partial load results the lower CO2 partial pressure in the calciner due to the flue gas recirculation, which increases the grid steam and oxygen partial pressures. The calcination model could overpredict the calcination reaction rate in the large-scale reactor. The average temperature in
In this section, the performance and behavior of the large scale calcium looping unit is analyzed in lower loads. In this context, the lower load means that the flue gas flow is lowered based on the
Table 3 Simulation parameters changed in different load scenarios. Parameter
Full load
75% load
50% load
30% load
0% load
Calciner inventory set (kg) Carbonator inventory (calculated) (kg) Flue gas flow (kg/s) Oxidant flow (kg/s) Fuel flow (kg/s) Calciner flue gas recirculation (kg/s) Carbonator flue gas recirculation (kg/s) Make-up flow (kg/s) Total cooling (MW)
35,091 50,398 116 26 9.23 82.13 0.0 11.15 58
35,067 49,548 87 19 6.92 59.00 0.0 8.36 51
35,050 47,039 58 18 6.20 58.00 28.90 5.57 34
35,064 29,220 35 16 5.90 60.00 52.09 3.34 31
35,042 19,135 0 15 5.50 59.00 85 2.00 30
335
J. Ylätalo et al. / Fuel 115 (2014) 329–337
gthermal ¼
Q cooling þ Q carb;gas þ Q calc;gas Q fuel
ð3Þ
Thermal power [MW]
Carbonation efficiency [%]
Total cooling [MW]
Thermal efficiency [%]
300
100 90
250 80 70 60 150
50 40
Efficiency [%]
Power [MW]
200
100 30 20 50 10 0
0 0
25
50
75
100
Flue gas load [%] Fig. 6. Thermal power of the calciner (left-hand axis), calcium looping unit cooling (left-hand axis) and capture efficiency (right-hand axis) plotted in different load scenarios.
Carbonator velocity [m/s]
Recirculation gas flow calciner [kg/s]
Calciner velocity [m/s]
Recirculation gas flow carbonator [kg/s]
6.5 80
Average velocity [m/s]
6
70
5.5
60
5
50 40
4.5 30 4
Gas mass flow [kg/S]
power of the combustor where it is originating from. For example, the 75% load means that the combustor is running at 75% and the carbonator flue gas flow is updated based on that. The zero load scenario means that the power plant is shut down and no flue gas is available for the carbonator. The purpose of this zero load simulation is to demonstrate that the calcium looping unit can be run as an independent oxy-combustion power generation unit or a back-up power plant if the calcium looping unit is independent from the original combustor. Table 3 lists the input changes in different load scenarios. All the other parameters were kept constant including the gas compositions and temperatures although they might change due to the changes in the combustor. The control scenarios have been devised based on the fluidizing conditions in the reactors. It was determined that achieving the CFB-mode becomes more difficult in the model when the fluidizing velocity drops below 4 m/s. From the 100% to the 75% load the adjustment can be done by means of lowering fluidizing velocities. Lower fluidizing velocity in the carbonator means lower solid circulation rate and the calciner fuel and oxygen flow can be adjusted accordingly. When the CFB mode is not achievable anymore in the carbonator solely with the flue gas flow, the incorporation of the flue gas recirculation is necessary. In the calciner this is already present to dilute the oxygen flow. With both reactors equipped with the wet flue gas recirculation, suitable fluidizing velocity can be achieved in different load scenarios. In the special case of the zero flue gas flow to the carbonator, the fluidization has to be handled with air. The flue gas recirculation has some positive side-effects, which are the increased CO2 and SO2 capture efficiencies in the carbonator. However this comes with a price, because sustaining fluidization also requires maintaining some thermal power, which can be seen in Fig. 6. This means that the relation of the calciner power to the combustor power increases in the lower loads, when in full power it is 53%, in the 30% load it is 70%. The thermal efficiency of loop increases for a bit in the lower loads because less heat needed for make-up calcination or heating up gas and solid flows. The thermal efficiency of the system was approximated from equation
20 3.5
10 0
3 0
50
100
Flue gas load [%] Fig. 7. Average velocities of the carbonator and calciner (left-hand axis) and recirculation of flue gas in the carbonator and calciner (right-hand axis) in different load scenarios.
where Qcooling is the heat extracted from carbonator and solid train (W), Qcarb,gas is the heat captured from the gas stream leaving the carbonator and Qcalc,gas is the heat captured from gas stream leaving the calciner. Total thermal power of the calciner is Qfuel. The exit temperature of the carbonator and calciner gas flows was selected to be 180 °C after the backpass heat exchangers to account for the dew point of sulfuric acid. This approach does not consider the performance of the backpass in lower gas flows. It very likely that the heat captured in the backpass is less in lower loads and the actual thermal efficiency is not higher in lower loads. The development of the flue gas recirculation is presented in Fig. 7. Below the 75% load the calciner flue gas recirculation is almost constant. In the carbonator the fuel gas recirculation increases linearly as load is decreased until in the zero load scenario the whole reactor is fluidized with air. In addition to that, Fig. 7 plots the average gas velocities in the reactors (left axis). The objective was to keep fluidization velocities constant in the reactors after the 75% load and use the solid material recirculation to adjust the flow of solids required for the CO2 capture. Nonetheless, the gas velocity drops in the calciner and increases in the carbonator due to increase in the reactor temperature differences, Fig. 8. Fig. 9 plots the adjustment of the make-up flow in different load scenarios and the control of solid flow between the reactors compared to the solid flow out of the carbonator. The make-up has been linearly controlled but it has a theoretical minimum set by the fuel ash and sulfur content. In the zero load case, the makeup flow does not have any significance in the CO2 capture, but it is needed to compensate the purge needed for the ash and CaSO4 removal. In the same plot the solid flow out of the carbonator is presented alongside the solid flow lead from carbonator to the calciner. Several factors affect the solid flow out of the carbonator. Below the 75% load, the attempt was to keep the gas velocity at around 4 m/s although the temperature change increased the gas velocity. This did not increase the solid flow because the carbonate content of the solids and the inventory of the carbonator decreases in the lower loads resulting in lower solid flows. The solid flow from the carbonator to the calciner was adjusted with solid recirculation, the percentage of recirculation had to be controlled case by case because the solid circulation rate out of the reactors is changing as a function of solid inventory and fluidizing velocity.
336
J. Ylätalo et al. / Fuel 115 (2014) 329–337
performance of the turbulent bed external heat exchanger in lower load is something that needs further investigation. Even if the flue gas load does not significantly change, several factors can affect the fluidization behavior of the calcium looping unit. Particle population of the system can change in size or density during the runs due to agglomeration or attrition which requires change in fluidization. Because the calcium looping process is sensitive to the solid circulation rates between the reactors, effective ways to control solid circulation are necessary. Even though there is some error in the prediction of solid circulation rates and heat transfer, the general observations from the simulations are applicable to the operation of large scale calcium looping units.
Average temperature of carbonator Average temperature of calciner
1000
900
Temperature [ºC]
800
700
4. Conclusion
600
500
400 0
25
50
75
100
Flue gas load [%] Fig. 8. The relation of exiting solids from the carbonator to the solid entering the calciner from the carbonator (left axis) and the make-up flow to the calciner (right axis) during different loads.
Solid flow from carbonator to calciner
Solid flow out of carbonator
Make-up flow 12
400
350 10
Solid mass flow [kg/s]
8 250
6
200
150 4 100
Make-up flow (CaCO3) [kg/s]
300
2
An existing 1-D dynamic calcium looping model was applied to a large scale calcium looping concept capturing CO2 from a 250 MWt power plant. Several new features were added to the model frame in order to successfully simulate the large scale unit, including new material fractions, ash and CaSO4, sulfur capture modeling and heat transfer in the solid return system. The plant was dimensioned based on the experience from large CFB units and the heat transfer design was evaluated based on the initial simulations of the design case. The unit was then simulated in five load scenarios ranging from full load to no flue gas flow to the carbonator. The results from the simulations confirmed that the successful operation of a large scale calcium looping unit requires good heat transfer design including the cooling of the hot solids coming from the calciner. Also the recirculation of flue gas in both reactors is necessary to ensure sufficient fluidization for different particle sizes or flue gas flows from the source combustor. Solid circulation control is also critical because it affects heavily the thermal balance of the system as well as the capture efficiency and CO2 balance. It was also observed that in the zero load situation fluidizing the carbonator with ambient air and running the calciner as an oxy-combustion circulating fluidized bed unit is also possible. If the power plant is abruptly shut down, the calcium loop can provide back-up power for some period. Utilizing the experience gained from large CFB-units coupled with new innovations, the scale-up of this process could be feasible in the near future. Further investigation of the solid mixing and temperature gradients in the carbonator and calciner bottom bed area with more intrinsic models like 3-D and CFD would confirm whether the capture efficiencies are as good as the 1-D model predicts.
50
Acknowledgements 0
0 0
25
50
75
100
Flue gas load [%] Fig. 9. Development of the reactor average temperatures during different loads.
An interesting observation is that in the zero load situation fluidizing the carbonator with ambient air and running the calciner as an oxy-combustion circulating fluidized bed unit is also possible. Minimal heat was extracted from the external heat exchanger and solid train surfaces and the rest can be extracted in backpasses of the reactors. If the power plant is abruptly shut down, the calcium loop can provide back-up power for some period. To summarize the different load scenario analysis, flexibility can be achieved using solid and flue gas recirculation and clever heat transfer design. The flue gas flow out of the calciner in the zero load case drops to 60% of the full load flow which will certainly have effect on the backpass heat exchanger performance. The thermal efficiency in the partial loads will not be as good as predicted by the simplified approach used in this study. Also the
The research leading to these results has received funding from the Carbon Capture and Storage Programme (2011–2015), financed by the Finnish Funding Agency for Technology and Innovation (Tekes) and coordinated by the Finnish Cluster for Energy and Environment (CLEEN Ltd.). References [1] Metz B, Davidson O, de Coninck H, Loos M, Meyer L. Special report on carbon dioxide capture and storage. In: Intergovernmental panel on climate change. Cambridge, (UK): Cambridge University Press; 2005. [2] Sánchez-Biezma A, Ballesteros JC, Diaz L, de Zárraga E, Álvarez FJ, López J, et al. Postcombustion CO2 capture with CaO. Status of the technology and next steps towards large scale demonstration. Energy Proc 2011;4:852–9. [3] Galloy A, Ströhle J, Epple B. Design and operation of a 1 MWth carbonate and chemical looping CCS test rig. VGB Power Tech 2011;91:64–8. [4] Hawthorne C, Dieter H, Bidwe A, Schuster A, Scheffknecht G, Unterberger S, et al. CO2 capture with CaO in a 200 kWth dual fluidized bed pilot plant. Energy Proc 2011;4:441–8. [5] Shimizu T, Hirama T, Hosoda H, Kitano K, Inagaki M, Tejima K. A twin fluid-bed reactor for removal of CO2 from combustion processes. Chem Eng Res Des 1999;77:62–8.
J. Ylätalo et al. / Fuel 115 (2014) 329–337 [6] Dean CC, Blamey J, Florin NH, Al-Jeboori MJ, Fennell PS. The calcium looping cycle for CO2 capture from power generation, cement manufacture and hydrogen production. Chem Eng Res Des 2011;89:836–55. [7] Martínes A, Lara Y, Lisbona P, Romeo LM. Energy penalty reduction in the calcium looping cycle. Int J Greenshouse Gas Control 2012;7:4–81. [8] Li Y, Liu C, Sun R, Liu H, Wu S, Lu C. Sequential SO2/CO2 capture of calciumbased solid waste from the paper industry in the calcium looping process. Ind. Eng. Chem. Res 2012;51:16042–8. [9] Rodríguez N, Murillo R, Abanades JC. CO2 capture from cement plants using oxyfired precalcination and/or calcium looping. Environ Sci Techonol 2012;46:2460–6. [10] Florin N, Fennell P. Synthetic CaO-based sorbent for CO2 capture. Energy Proc 2011;4:830–8. GHGT-10. [11] Martínez I, Murillo R, Grasa G, Rodríguez N, Abanades JC. Conceptual design of a three fluidised beds combustion system capturing CO2 with CaO. Int J Greenhouse Gas Control 2011;5:498–504. [12] Nikolopoulos A, Nikolopoulos N, Charitos A, Grammelis P, Kakaras E, Bidwe AR, et al. High-resolution 3-D full-loop simulation of a CFB carbonator cold model. Chem Eng Sci 2013;90:137–50. [13] Vorrias I, Atsonios K, Nikolopoulos A, Nikolopoulos N, Grammelis P, Kakaras E. Calcium looping for CO2 capture from a lignite fired power plant. Fuel 2013, In press http://dx.doi.org/10.1016/j.fuel.2012.12.087. [14] Hawthorne C, Trossmann M, Cifre PG, Schuster A, Scheffknecht G. Simulation of the carbonate looping power cycle. Energy Procedia Greenhouse Gas Control Technologies 9. In: Proceedings of the 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), 16–20 November 2008, Washington, DC, USA, 2009;1:1387–94. [15] Abanades JC, Anthony EJ, Wang J, Oakey JE. Fluidized bed combustion systems integrating CO2 capture with CaO. Environ Sci Technol 2006;39:2861–6. [16] Romano M. Coal-fired power plant with calcium oxide carbonation for postcombustion CO2 capture. Energy Proc 2009;1:1099–106.
337
[17] Diego ME, Arias B, Abanades JC. Modeling the solids circulation rates and solids inventories of and interconnected circulating fluidized bed reactor system for CO2 capture by calcium looping. Chem Eng J 2012;198–199:228–35. [18] Martínez I, Murillo R, Grasa G, Abanades JC. Integration of a Ca-Looping system for CO2 capture in an existing power plant. Energy Proc 2011;4:1699–706. [19] Ylätalo J, Ritvanen J, Arias B, Tynjälä T, Hyppänen T. 1-Dimensional modeling and simulation of the calcium looping process. Int J Greenhouse Gas Control 2012;9:130–5. [20] Johnsson F, Leckner B. Vertical distribution of solids in a CFB-furnace, In: Heinschel K, editor, Proceedings of the 13th International Conference on Fluidized Bed combustion Orlando, FL, USA. May 07–10, 1995. p. 671–679. [21] Fang F, Li Z, Cai N. Experiment and modeling of CO2 capture from flue gases at high temperature in a fluidized bed reactor with Ca-based sorbents. Energy Fuels 2009;23:207–16. [22] Myöhänen K, Hyppänen T. A three-dimensional model frame for modelling combustion and gasification in circulating fluidized bed furnaces. Int J Chem Reactor Eng 2011;9. Article A25. [23] Alonso M, Rodríguez N, Gonzáles B, Grasa G, Murillo R, Abanades JC. Carbon dioxide capture from combustion flue gases with a calcium oxide chemical loop. Experimental results and process development. Int J Greenhouse Gas Control 2010;4:167–73. [24] Ylätalo J, Parkkinen J, Ritvanen J, Tynjälä T, Hyppänen T. Modeling of the oxycombustion calciner in the post-combustion calcium looping process. Fuel 2012. In press. http://dx.doi.org/10.1016/j.fuel.2012.11.041. [25] De-Souza Santos M. Solid fuels combustion and gasification. CRC Press; 2010. ISBN-9781420047493. [26] González B, Alonso M, Abanades JC. Sorbent attrition in a carbonation/ calcination pilot plant for capturing CO2 from flue gases. Fuel 2010;89:2918–24.
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Model based scale-up study of the calcium looping process Jaakko Ylätalo ⇑, Jouni Ritvanen, Tero Tynjälä, Timo Hyppänen Lappeenranta University of Technology, LUT Energy, P.O. Box 20, 53851 Lappeenranta, Finland
h i g h l i g h t s Post-combustion calcium looping unit was retrofitted for a 250 MWth combustor. Dimensioning of the unit was done with a 1-D dynamic model. Five flue gas load cases were analysed from 100% to 0% load. Flue gas and solid circulation played a key role in control of the system. Heat transfer and solid transfer design are important in large scale units.
a r t i c l e
i n f o
Article history: Received 17 June 2013 Received in revised form 10 July 2013 Accepted 11 July 2013 Available online 25 July 2013 Keywords: Calcium looping process 1-D modeling CO2 capture Fluidized bed reactor Scale-up
a b s t r a c t A 1-D dynamic calcium looping model was applied to a large scale calcium looping concept capturing CO2 from a 250 MWt power plant. Several new features were added to the existing model frame in order to successfully simulate the large scale unit. Models were needed for new material fractions, such as ash and CaSO4, sulfur capture and heat transfer in the solid return system. The plant was dimensioned based on the experience from large CFB units and the heat transfer design was evaluated based on initial simulations of the design case. The unit was then simulated in five load scenarios ranging from full load to zero load, no flue gas flow to the carbonator. The scale-up of the process is feasible from the model’s point of view keeping in mind the assumptions and simplifications made in the modeling. The results from the simulations confirmed that successful operation of a large scale calcium looping unit requires good heat transfer design including cooling of the hot solids coming the calciner. Also the recirculation of flue gas in both reactors is necessary to ensure the sufficient fluidization for different flue gas flows from the source combustor. Solid circulation control is also critical because it affects heavily the thermal balance of the system as well as the capture efficiency and the CO2 balance. Results show that utilizing the experience gained from the large CFB-units coupled with new innovations, the scale-up of this process could be feasible in the near future. The ability to operate flexibly in different modes could give this technology an advantage needed for wide industrial utilization of the process. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The scientific community around the world is almost unanimous that the actions of mankind have accelerated the release of greenhouse gases into the world’s atmosphere and consequently increased the chance of rapid climate change in the coming decades. The general opinion among legislators and the scientific community is that swift actions should be taken to prevent global warming by cutting down greenhouse gas emissions [1]. It is apparent that this cannot be achieved by single actions but by applying a variety of methods ranging from clean energy production to population and economy growth control. Energy production is one of the largest contributors to greenhouse gas emissions but
⇑ Corresponding author. Tel.: +358 40 171 4003. E-mail address: jaakko.ylatalo@lut.fi (J. Ylätalo). 0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2013.07.036
it is also approachable from a technical perspective so the potential to cut down emissions quickly is high. The demand for energy is increasing year by year and actions should be taken to reduce the emissions from fossil fuels. This can be done by utilizing a variety of methods simultaneously, carbon capture and storage as one of the most promising methods to cut down emissions quickly from existing stationary sources. Several carbon capture methods are being developed to answer the need to reduce stationary emissions from fossil fuel power plants. Oxy-combustion and amine-based post-combustion capture methods are technologically more mature and in the demonstration phase. Second generation capture technologies like chemical looping and calcium looping are also developing quickly and are advancing in the pilot stage [2–4]. Calcium looping is a technology that uses lime to absorb CO2 from flue gases in a fluidized bed reactor and release it in a second fluidized bed regenerator to produce a clean CO2 flow
330
J. Ylätalo et al. / Fuel 115 (2014) 329–337
to the storage [5,6]. Post-combustion calcium looping and all of its variations [7–9] are promising technologies for reducing CO2 emissions from coal-fired power plants because the availability of limestone and potential for high CO2 and SO2 capture rate. Also the technological advancements in calcium looping, like thermal coupling of the combustor and carbonator or high performance sorbents, can increase process efficiency and greatly reduce the energy penalty due to carbon capture. [10,11]. One of the major issues in the second generation capture technologies is the scale-up to industrial scale units. The construction of these industrial units requires a large capital investment. However, for the calcium looping technology to succeed, operational experience from large scale units is required, before new technologies can be utilized in commercial scale. This can be partly achieved by using a variety of computational tools available for process modeling and simulation. Utilizing a variety of tools, different phenomena can be researched, for example, CFD calculations enable the study of gas–solid interactions and dual fluidized bed behavior but are computationally very strenuous [12]. Process modeling tools like AspenÒ or IPSEproÒ help to understand the whole process and how different phenomena affect each other and are computationally more forgiving but rely on several assumptions [13–15]. Both approaches are required to comprehensively understand the process and gradually transfer systems to the industrial scale. In this paper, a modeling approach is used to scale-up the postcombustion calcium looping process to capture CO2 from a conceptual 250 MWt power plant. A 1-D dynamic process model is used to dimension the reactors and simulate the interconnected system ensuring that the operational parameters are feasible and to study the behavior of the unit in different load scenarios. Several studies have been published [16–18] where large scale calcium looping systems have been simulated but none of the studies has incorporated reactor design, thermal integration and overall loop simulation in one study. 2. Modeling approach The calcium looping process is based on the partially reversible reaction between calcium oxide and carbon dioxide. CO2 from the flue gases of a combustor are absorbed in a fluidized bed of lime, carbonator. Entrained solids containing CaCO3 are transferred to a regenerator, calciner, where CO2 is released from the solid material. Regenerated CaO is entrained back to the carbonator where the loop is closed. A 1-D dynamic calcium looping model will be applied to this looping process [19]. The 1-D model will be used to study the dimensioning and operation of the reactor units to determine whether there are any technical or engineering problems in the scale-up. 2.1. The 1-D calcium looping model The 1-D dynamic calcium looping model frame has been built piece by piece adding important phenomena and coupling reactors and loop components. The model frame is solved in the MatlabÒ Simulink environment using built-in ordinary differential equation solvers. The model is capable of both dynamic and steady-state simulations. The carbonator and calciner are spatially discretized into 1-D control volumes in which mass and energy balances are calculated each time step using the first order upwind scheme. The reactors are interconnected in the Simulink environment and 0-D models of the loop seals, cyclones and standpipes exist in the model frame. The reactor model frames incorporate several fluidized bed phenomena like solid behavior in a circulating fluidized bed, solid suspension heat transfer, core-annulus calculation, solid
fuel combustion, sulfur capture, ash accumulation and cyclic carbonation and calcination. For solid circulation and suspension, semi-empirical models have been applied from the literature. The vertical division of solids in the reactor is modeled with a modified approach presented by Johnsson and Leckner [20]. The equation was modified by leaving out the static bed term. Solid flow out of the reactor is modeled by a semi-empirical equation including exit density of the reactor, slip velocity of solids in the exit and exit channel dimensions [19]. Both the solid density profile and solid flow out of the reactor are tied to the total solid inventory of the reactor, particle properties and gas velocity profile of the reactor. The rates of chemical reactions have been defined from the best available knowledge from the literature. The reaction rate for carbonation was applied from the work of Shimitzu et al. [5]. Carbonation is limited by the aging of the lime by defining maximum carbonate content for the solids. Calcination reaction rate was acquired from the model proposed by Fang et al. [21]. The combustion model has been built based on an approach presented by Myöhänen and Hyppänen [22]. The combustion model divides the input fuel into char, volatiles, ash and moisture. Fuel moisture is released immediately to the gas balance. A time constant can be defined for the release of volatiles. Volatiles are combusted in the gas balance based reaction rates acquired from the literature. Char and ash are divided into char balance and solid balance. Char combustion is modeled with a simple model homogeneous combustion model. Ash is considered inert and will be a part of the system solids with homogeneous properties. The core-annulus model enables the transfer of material and heat from the top part of the reactor to the bottom part but the material in the annulus is not considered reacting. Mass and energy balances for the core and annulus are solved separately. Solids flow in the core upwards and downwards in the annulus. The coreannulus model can be used to emulate the internal flow of solids in a fluidized bed reactor. The amount of solid flow in the annulus is determined based on experimental results keeping in mind that the solid volume fraction cannot exceed 0.6 and the solid velocity in the core cannot exceed gas velocity. In addition to the coreannulus approach, mixing of energy in consecutive control volumes can be modeled with a turbulent diffusion. The energy balance incorporates several phenomena present in a fluidized bed reactor. In addition to energy transported alongside solid and gas flows, heat losses through the walls can be modeled as well as heat transfer to membrane walls with or without refractory. Single heat sinks or sources can be defined in each control volume. Heat produced or consumed by chemical reactions is also included in each control volume energy balance. The model has been validated both with lab-scale equipment as well as with 1.7 MW pilot scale experimental results [2,23]. Alongside that, validation and comparison between the 1-D and a 3-D steady-state calciner model has been done [24]. Generally, the maturity of the model starts to near a tool suitable for scale-up purposes. The 1-D model tool can be used to evaluate the solid circulation between the reactors, the amount of heat transfer surface required inside the reactors, performance in different loads, the amount of make-up required for successful CO2 and SO2 capture without performance loss and many other features. Studying these features will point out whether the calcium loop is technically feasible for the calculated case. 2.2. Improvements and updates to the model frame The model frame has been improved since the last publications [19,24] by adding several features to more accurately describe the process. The calculation of material fractions CaSO4 and ash has been added to the solid balance in order to model sulfation and ash accumulation. The equations to calculate the solid material
J. Ylätalo et al. / Fuel 115 (2014) 329–337
fractions remain the same as in Ylätalo et al. [19], only now instead of two materials, CaO and CaCO3, there are four. The sulfation rate equation has been implemented from De-Souza Santos [25] taking into account only indirect sulfation, the reaction between calcium oxide and sulfur dioxide. This assumption is made because the calciner has an abundance of CaO for sulfur capture compared to CaCO3 even though the calciner can operate near the equilibrium. 8810 T
r sulf ;CaO ¼ 4:9 103 ð3:843T þ 5640Þe
X CaO xSO2 xO2
ð1Þ
where rsulf,CaO is the sulfation rate of calcium oxide (1/s), T is the surrounding temperature (°C), XCaO is the calcium oxide molar fraction (-), xSO2 and xO2 are the molar fractions of SO2 and O2 in the surrounding gas. Tracking material fractions can also be used to design solid material purge and make-up flows for the system. Another feature is the ability to extract heat from the solid return system by using simple heat balance calculations in the loop seal, stand pipe and cyclone computational blocks. This addition enables better thermal design of the system which will be demonstrated in the following sections. The solved equation for the solid return system is
331
Table 1 Basic information of the 250 MW simulation case and some essential simulations parameters. Combustor thermal power (MW) Flue gas mass flow (kg/s) Flue gas CO2 (wt%) Flue gas SO2 (wt%) Flue gas temperature (°C) Designed capture efficiency (%) Calciner fuel flow (kg/s) Fuel LHV (MJ/kg) Oxidant flow (kg/s) Make-up flow (kg/s) Estimated maximum average CaCO3 fraction (wt%) Fuel char (wt%) Fuel volatiles (wt%) Fuel moisture (wt%) Fuel ash (wt%) Average particle size (lm) Average particle density (kg/m3) Turbulent dispersion coefficient (-) Circulation coefficient (-) Wall layer flow speed (m/s)
250 115.92 21.16 0.17 120 80 9.229 30 26 11.15 41 67.6 22.5 7.5 2.4 200 1800 0.3 0.3 0.01
Q removed ¼ qm;s cp;s ðT in;s T element Þ ¼ hin Awall ðT element T wall Þ ¼
kwall Awall ðT wall T out Þ swall
ð2Þ
The thermal power removed from the system is represented by Qremoved (W), qm,s is the solid flow through the system (kg/s), cp,s is the solid heat capacity (J/kgK), Tin,s is the solid temperature entering (°C) and Tblock is the internal temperature of the 0-D block (°C) referring for example to the loop seal internal temperature. The heat transfer coefficient of the block is hin, approximately 100– 200 W/m2K, Awall is the inside wall area or the block and Twall is the wall temperature. The heat conductivity of the wall material is kwall (W/mK) and swall (m) is the thickness of the material and Tout is the outer temperature of the block (°C) for example heat transfer surface or outside metal temperature. The loop seal, standpipe and cyclone can be insulated with refractory or they can be equipped with heat transfer surfaces. A PID-controller has been constructed to control the calciner purge to handle the total solid inventory of the system more efficiently and to limit the accumulation of ash and CaSO4. PID-controller follows the calciner inventory and changes the purge according to the desired calciner inventory defined in the input. Make-up of the system is kept constant. The controller slows down the simulation significantly but increases stability which is an acceptable trade-off. The PID-controller resembles the actions of an operator who would follow the pressure difference of the bed and adjust the calciner purge. Further analysis of the system control is not done in this study: the controller is used to find a stable solution faster than by iterating the purge by hand, although the control loop increases calculation times. 3. Results 3.1. Introduction of the scale-up case A conceptual power plant with a thermal power of 250 MW was selected as the scale-up case. Several factors affected the size of the selected case. The pre-industrial size poses less financial risk but can be successfully extrapolated to a larger scale. Also the retrofitting with the calcium loop doubles the thermal output of the plant. General operation parameters of the power plant and the calcium loop are listed in Table 1 as well as the composition of the primary fuel and some essential simulation parameters. Currently the model can handle only one fuel so high quality coal is selected for the combustor and the calcium loop although power plants tend to use several fuels in combination. The calcium looping unit was
modeled as a retrofit unit without steam cycle or thermal integration to the 250 MW boiler.
3.2. Dimensioning and thermal design of the reactors The flue gas flow rate of the carbonator can be used to dimension the carbonator reactor once the average particle size is determined for the system. For the selected fuel and air-ratio of 1.2, the flue gas flow is 116 kg/s. Studies have shown that calcium looping systems tend to have a small particle size (<100 lm) due to fragmentation in the initial calcination [26]. With this particle size, it is quite difficult to operate in the conventional CFB-region. That is why it is more practical to use a higher average particle size commonly found in CFB-units. If the average particle size is fixed to 200 lm, the maximum superficial gas velocity can be set to 6 m/s. Both particle sizes are in the Geldart A class but the 100 lm particles approach the pneumatics transport region. A larger particle size could be achieved by using more durable limestone against fragmentation or increasing initial feed size. Calculating from the carbonator gas mass flow, approximate density and the maximum velocity, the cross section of the carbonator will be around 44 m2. Because the gas velocity in the carbonator is decreasing due to the adsorption of CO2 by the CaO, the use of sloped bottom section, commonly found in CFB boilers, is not appropriate. Also using secondary flue gas inputs is not necessary because that will only cause some of the flue gas CO2 to bypass the dense, reactive part of the carbonator bed. The rectangle like cross section commonly used in CFB-boilers is not necessary but in order to keep structural integrity with the calciner, the carbonator width is determined to be the same as the calciner. The height of both reactors could be defined to maximize the residence time of gas and solids but in practice, the height the cyclone, standpipe and loop seal combination sets the maximum height of the system. In this case, a value of 35 m could be close to the reality when examining existing CFB boilers. In the thermal design it was determined that an external heat exchanger would be the best solution to remove heat from the incoming solids from the calciner and carbonation reaction. An evaporator surface in the carbonator would be inflexible in terms of cooling in lower loads. Initial simulations also pointed out that, if the carbonator bed is refractory protected due to the very high erosion rate of the dense bed, the heat transfer rate in the bed is not enough to cool the solids to 650 °C. Using in-bed heat transfer tubes would quickly cool the solids but it would also lead to fast
332
J. Ylätalo et al. / Fuel 115 (2014) 329–337
Fig. 1. Concept image of the large calcium looping unit with dimensions and material flows.
erosion of the tubes. The most reactive bed part would be outside of the equilibrium which would cause trouble in the capture. It was determined that the cooling the solids before feeding them into the carbonator would ease the heat transfer load in the carbonator bottom part. In addition, the equipment feeding the solids are also under great thermal stress and cooling of them is necessary. Location of external heat exchanger in the solid train feeding the carbonator is presented in Fig. 1. In the model the external heat exchanger is executed by cooling the incoming solids to a low temperature (550–600 °C). In an actual unit, the external heat exchanger could receive solids from the carbonator to cool the reactor more efficiently. Calciner dimensions can be evaluated similarly keeping in mind that the calciner gas flow consists of the oxygen required for combustion, CO2 from the recirculation gas and the CO2 released from the calcination of CaCO3 in the circulated material and make-up flow. A superficial velocity of 6 m/s in the freeboard was selected as a boundary condition and the fuel flow was defined from a simple 0-D balance, with these assumptions the gas flow out of the calciner is around 108 kg/s. When the operation temperature is around 950 °C, the freeboard cross section is 60 m2. The primary gas flow can be adjusted by reducing the amount of recirculation gas but it has to be kept in mind that the amount of oxygen in the gas should not rise over 40 vol%, which could lead to too high combustion temperatures. In the calciner, a sloped section is necessary because the difference between the freeboard and grid gas mass flows is significant. By evaluating the grid gas mass flow and the exiting gas flow relation and the average time taken to reach the gas flow to fully develop, it was determined that the sloped section will be 5 m high and the grid cross section will be 43 m2. With this dimensioning, the calciner velocity should remain constant along the reactor height. The width of the calciner was set to 3.8 m to keep the fuel, gas and solid penetration sufficient. Rectangular cross section commonly used in CFB-units was also selected for the calciner. This will set the length of the reactor to 15.8 m and the width of the grid section, 2.7 m, because the length is not changed in the grid. The carbonator width will be 3.8 m and depth 11.5 m and both reactors will be structurally connected at the ends, Fig. 1. This structural connection enables the construction of the solid return system alongside the reactor and the feeding of a portion of the solids back to the original reactor and transferring the remaining solids to the next reactor. The wall between the
reactors is insulated to prevent heat transfer between reactors. Also this solid train type of solution enables the division of solid feeding along the reactor length to even out the active solid concentration in the reactor. This kind of solution can be modularized: larger units could be built by adding consecutive carbonator and calciner blocks, increasing the length. The main dimensions of the reactors and the thermal design solutions are listed in Table 2. Both reactors are fitted with the ability to recirculate solids and gas. The reason for this is explained in the following sections. The calciner is fluidized with the mixture of recirculation gas and oxygen. The carbonator can be fluidized solely with flue gas from the combustor or with a combination of recirculation gas from the carbonator or just with air. 3.3. Full load results The simulation results for the full load case and design point of the plant can be seen in Figs. 2–5. Fig. 2 describes the overall balance of the system. The make-up flow for the case was determined based on the fuel ash and sulfur content, achieving an average activity of 41 wt% in the system. With the 6 m/s fluidization velocity, the solid mass flow out of the reactors is much larger than needed for moderate capture therefore 40% of the solids is recirculated in both reactors. The fluidization velocity of the calciner is purposefully lower than the carbonator velocity because it helps to control the overall solid inventory of the system. If the calciner velocity becomes higher and solids start to move to the carbonator, there is a danger that the whole reactor is drained from solids.
Table 2 Main inner dimensions of the carbonator and calciner reactors and cooling surface parameters of the large scale calcium looping plant. Dimension (inner)
Carbonator
Calciner
Height of reactor (m) Cross section of grid (m2) Cross section of freeboard (m2) Width of grid (m) Width of freeboard (m) Length (m) End of sloped section height (m) Elevation to exit channels (m) Thermal design
35 44 44 3.8 3.8 11.5 None 33 External HE
35 43 60 2.7 3.8 15.8 5 33 Insulated
333
J. Ylätalo et al. / Fuel 115 (2014) 329–337
Exit gas flow [kg/s]
95
Exit gas CO2 [w-%]
3.88
Exit gas flow [kg/s] Exit gas CO2 [w-%]
141
Exit gas O2 [w-%]
4.39
Exit gas O2 [w-%]
Exit gas N2 [w-%]
87.67
Exit gas N2 [w-%]
0.20
88.66 3.76
Exit gas H2O [w-%]
4.07
Exit gas H2O [w-%]
7.22
Exit gas SO2 [w-%]
0.00
Exit gas SO2 [w-%]
0.00
Exit gas T [°C]
654
Exit gas T [°C] Carbonator solid exit
Inventory [kg] Inventory [kg/m2]
50398 1145
Calciner solid exit
Solid mass flow out [kg/s]
362
Solid mass flow out [kg/s]
Solid flux out [kg/m2 s]
8.2
Solid flux out [kg/m2 s]
CaO [w-%]
43.7
CaO [w-%]
299 5.0
Inventory [kg]
78.2
Inventory [kg/m2]
Ecarb [%]
85
CaCO3 [w-%]
37.9
CaCO3 [w-%]
0.1
Ecalc [%]
ESO2 [%]
100
CaSO4 [w-%]
9.6
CaSO4 [w-%]
11.2
ESO2 [%]
Xave [w-%]
0.41
Ash [w-%]
8.8
Ash [w-%]
10.5
Fuel mass flow [kg/s]
Recirculation [-]
0.4
Recirculation [-]
0.4
Make-up [kg/s]
Solid flow between [kg/s]
217
Solid flow between [kg/s]
179
Purge [kg/s]
Total cooling [MW]
Flue gas flow [kg/s] Flue gas CO2 [w-%]
58
894
116
Primary gas flow [kg/s]
21.16
Primary gas CO2 [w-%]
35091 584 99 98 9.23 11.15 8.94
26 0.00
Flue gas O2 [w-%]
3.63
Primary gas O2 [w-%]
100.00
Flue gas N2 [w-%]
71.72
Primary gas N2 [w-%]
0.00
Flue gas H2O [w-%]
3.33
Primary gas H2O [w-%]
0.00
Flue gas SO2 [w-%]
0.17
Primary gas SO2 [w-%]
0.00
Flue gas T [°C]
120
Primary gas T [°C]
Recirculation [kg/s]
0.00
Recirculation [kg/s]
82.13
150
Recirculation T [°C]
120
Recirculation T [°C]
120
Fig. 2. Collected 0-D results of the full load simulation case.
35 35
30
30
25
Height [m]
25
20
20
15
15
10
10
5
5
0 0.00
0.10
0.20
Solid volume fraction [-]
0.30
0 550
600
650
700
750
Temperature [°C]
Fig. 3. 1-D plots of the carbonator solid volume fraction and temperature in the full load case.
When the purge and solid control is on the calciner side, it is beneficial to keep the velocity difference favoring the calciner, Fig. 5. Fig. 3 displays the solid volume fraction profile and temperature profile of the carbonator. Especially from the temperature profile we can see that the temperature conditions are far from ideal even in the 1-D model. The entering solids are cooler than the objective temperature of the reactor. Total thermal power removed from the external heat exchanger and solid train is 58 MW. Fig. 4 presents the corresponding solid volume fraction and temperature profiles
for the calciner. Comparing the solid volume fraction profiles, it can be seen that the carbonator has a higher inventory than the calciner which could be a result of the higher CaCO3 content in the carbonator. The CaCO3 concentration in the solids is quite close to the average activity achievable because recirculation increases the residence time of the solids in the reactor. Another interesting observation is that the calciner is running in lower temperatures than the literature suggests. However, very good calcination efficiency is achieved. This could be a result of
J. Ylätalo et al. / Fuel 115 (2014) 329–337
Height [m]
334 35
35
30
30
25
25
20
20
15
15
10
10
5
5
0 0.00
0.10
0.20
0.30
0 800
Solid volume fraction [-]
850
900
950
1000
Temperature [°C]
Fig. 4. 1-D plots of the calciner solid volume fraction and temperature in the full load case.
Calciner
Carbonator 35
30
Height [m]
25
20
15
10
5
0 3.000
4.000
5.000
6.000
7.000
Gas velocity [m/s] Fig. 5. 1-D velocity profiles of the carbonator and calciner reactors in the full load case. The calciner velocity is lower than the designed 6 m/s because the temperature is lower than originally intended especially in the lower section of the reactor.
the 1-D control volumes is quite close to the equilibrium temperature of the calcination reaction, which could cause some problems in a real unit because of the lateral temperature profiles present in a real large scale unit. In the temperature profile the insertion point of the cold solids can be seen as a drop of the temperature in the bottom region. Also the fuel insertion point close to the height of 5 m increases the local temperature by 50 K. The calcination reaction is temperature driven, which means that in the calciner the solid material has a long time to react compared to the carbonator, in which it is critical to achieve good gas–solid contact in the dense bottom region of the reactor. In other words, the calciner operation is not so sensitive to local areas of temperatures below the calcination temperature in the lower part of the reactor. The performance of the system is higher than originally intended. The capture efficiency is 85%, when the design value was 80%. The sulfur capture is also almost 100% in both reactors. Fuel flow of the calciner has been evaluated to 9.23 kg/s which is equal to 277 MW thermal power. The ratio of the calciner thermal power from total thermal power, Qcalciner/(Qcalciner + Qcombustor), is around 53% in full load.
3.4. Partial load results the lower CO2 partial pressure in the calciner due to the flue gas recirculation, which increases the grid steam and oxygen partial pressures. The calcination model could overpredict the calcination reaction rate in the large-scale reactor. The average temperature in
In this section, the performance and behavior of the large scale calcium looping unit is analyzed in lower loads. In this context, the lower load means that the flue gas flow is lowered based on the
Table 3 Simulation parameters changed in different load scenarios. Parameter
Full load
75% load
50% load
30% load
0% load
Calciner inventory set (kg) Carbonator inventory (calculated) (kg) Flue gas flow (kg/s) Oxidant flow (kg/s) Fuel flow (kg/s) Calciner flue gas recirculation (kg/s) Carbonator flue gas recirculation (kg/s) Make-up flow (kg/s) Total cooling (MW)
35,091 50,398 116 26 9.23 82.13 0.0 11.15 58
35,067 49,548 87 19 6.92 59.00 0.0 8.36 51
35,050 47,039 58 18 6.20 58.00 28.90 5.57 34
35,064 29,220 35 16 5.90 60.00 52.09 3.34 31
35,042 19,135 0 15 5.50 59.00 85 2.00 30
335
J. Ylätalo et al. / Fuel 115 (2014) 329–337
gthermal ¼
Q cooling þ Q carb;gas þ Q calc;gas Q fuel
ð3Þ
Thermal power [MW]
Carbonation efficiency [%]
Total cooling [MW]
Thermal efficiency [%]
300
100 90
250 80 70 60 150
50 40
Efficiency [%]
Power [MW]
200
100 30 20 50 10 0
0 0
25
50
75
100
Flue gas load [%] Fig. 6. Thermal power of the calciner (left-hand axis), calcium looping unit cooling (left-hand axis) and capture efficiency (right-hand axis) plotted in different load scenarios.
Carbonator velocity [m/s]
Recirculation gas flow calciner [kg/s]
Calciner velocity [m/s]
Recirculation gas flow carbonator [kg/s]
6.5 80
Average velocity [m/s]
6
70
5.5
60
5
50 40
4.5 30 4
Gas mass flow [kg/S]
power of the combustor where it is originating from. For example, the 75% load means that the combustor is running at 75% and the carbonator flue gas flow is updated based on that. The zero load scenario means that the power plant is shut down and no flue gas is available for the carbonator. The purpose of this zero load simulation is to demonstrate that the calcium looping unit can be run as an independent oxy-combustion power generation unit or a back-up power plant if the calcium looping unit is independent from the original combustor. Table 3 lists the input changes in different load scenarios. All the other parameters were kept constant including the gas compositions and temperatures although they might change due to the changes in the combustor. The control scenarios have been devised based on the fluidizing conditions in the reactors. It was determined that achieving the CFB-mode becomes more difficult in the model when the fluidizing velocity drops below 4 m/s. From the 100% to the 75% load the adjustment can be done by means of lowering fluidizing velocities. Lower fluidizing velocity in the carbonator means lower solid circulation rate and the calciner fuel and oxygen flow can be adjusted accordingly. When the CFB mode is not achievable anymore in the carbonator solely with the flue gas flow, the incorporation of the flue gas recirculation is necessary. In the calciner this is already present to dilute the oxygen flow. With both reactors equipped with the wet flue gas recirculation, suitable fluidizing velocity can be achieved in different load scenarios. In the special case of the zero flue gas flow to the carbonator, the fluidization has to be handled with air. The flue gas recirculation has some positive side-effects, which are the increased CO2 and SO2 capture efficiencies in the carbonator. However this comes with a price, because sustaining fluidization also requires maintaining some thermal power, which can be seen in Fig. 6. This means that the relation of the calciner power to the combustor power increases in the lower loads, when in full power it is 53%, in the 30% load it is 70%. The thermal efficiency of loop increases for a bit in the lower loads because less heat needed for make-up calcination or heating up gas and solid flows. The thermal efficiency of the system was approximated from equation
20 3.5
10 0
3 0
50
100
Flue gas load [%] Fig. 7. Average velocities of the carbonator and calciner (left-hand axis) and recirculation of flue gas in the carbonator and calciner (right-hand axis) in different load scenarios.
where Qcooling is the heat extracted from carbonator and solid train (W), Qcarb,gas is the heat captured from the gas stream leaving the carbonator and Qcalc,gas is the heat captured from gas stream leaving the calciner. Total thermal power of the calciner is Qfuel. The exit temperature of the carbonator and calciner gas flows was selected to be 180 °C after the backpass heat exchangers to account for the dew point of sulfuric acid. This approach does not consider the performance of the backpass in lower gas flows. It very likely that the heat captured in the backpass is less in lower loads and the actual thermal efficiency is not higher in lower loads. The development of the flue gas recirculation is presented in Fig. 7. Below the 75% load the calciner flue gas recirculation is almost constant. In the carbonator the fuel gas recirculation increases linearly as load is decreased until in the zero load scenario the whole reactor is fluidized with air. In addition to that, Fig. 7 plots the average gas velocities in the reactors (left axis). The objective was to keep fluidization velocities constant in the reactors after the 75% load and use the solid material recirculation to adjust the flow of solids required for the CO2 capture. Nonetheless, the gas velocity drops in the calciner and increases in the carbonator due to increase in the reactor temperature differences, Fig. 8. Fig. 9 plots the adjustment of the make-up flow in different load scenarios and the control of solid flow between the reactors compared to the solid flow out of the carbonator. The make-up has been linearly controlled but it has a theoretical minimum set by the fuel ash and sulfur content. In the zero load case, the makeup flow does not have any significance in the CO2 capture, but it is needed to compensate the purge needed for the ash and CaSO4 removal. In the same plot the solid flow out of the carbonator is presented alongside the solid flow lead from carbonator to the calciner. Several factors affect the solid flow out of the carbonator. Below the 75% load, the attempt was to keep the gas velocity at around 4 m/s although the temperature change increased the gas velocity. This did not increase the solid flow because the carbonate content of the solids and the inventory of the carbonator decreases in the lower loads resulting in lower solid flows. The solid flow from the carbonator to the calciner was adjusted with solid recirculation, the percentage of recirculation had to be controlled case by case because the solid circulation rate out of the reactors is changing as a function of solid inventory and fluidizing velocity.
336
J. Ylätalo et al. / Fuel 115 (2014) 329–337
performance of the turbulent bed external heat exchanger in lower load is something that needs further investigation. Even if the flue gas load does not significantly change, several factors can affect the fluidization behavior of the calcium looping unit. Particle population of the system can change in size or density during the runs due to agglomeration or attrition which requires change in fluidization. Because the calcium looping process is sensitive to the solid circulation rates between the reactors, effective ways to control solid circulation are necessary. Even though there is some error in the prediction of solid circulation rates and heat transfer, the general observations from the simulations are applicable to the operation of large scale calcium looping units.
Average temperature of carbonator Average temperature of calciner
1000
900
Temperature [ºC]
800
700
4. Conclusion
600
500
400 0
25
50
75
100
Flue gas load [%] Fig. 8. The relation of exiting solids from the carbonator to the solid entering the calciner from the carbonator (left axis) and the make-up flow to the calciner (right axis) during different loads.
Solid flow from carbonator to calciner
Solid flow out of carbonator
Make-up flow 12
400
350 10
Solid mass flow [kg/s]
8 250
6
200
150 4 100
Make-up flow (CaCO3) [kg/s]
300
2
An existing 1-D dynamic calcium looping model was applied to a large scale calcium looping concept capturing CO2 from a 250 MWt power plant. Several new features were added to the model frame in order to successfully simulate the large scale unit, including new material fractions, ash and CaSO4, sulfur capture modeling and heat transfer in the solid return system. The plant was dimensioned based on the experience from large CFB units and the heat transfer design was evaluated based on the initial simulations of the design case. The unit was then simulated in five load scenarios ranging from full load to no flue gas flow to the carbonator. The results from the simulations confirmed that the successful operation of a large scale calcium looping unit requires good heat transfer design including the cooling of the hot solids coming from the calciner. Also the recirculation of flue gas in both reactors is necessary to ensure sufficient fluidization for different particle sizes or flue gas flows from the source combustor. Solid circulation control is also critical because it affects heavily the thermal balance of the system as well as the capture efficiency and CO2 balance. It was also observed that in the zero load situation fluidizing the carbonator with ambient air and running the calciner as an oxy-combustion circulating fluidized bed unit is also possible. If the power plant is abruptly shut down, the calcium loop can provide back-up power for some period. Utilizing the experience gained from large CFB-units coupled with new innovations, the scale-up of this process could be feasible in the near future. Further investigation of the solid mixing and temperature gradients in the carbonator and calciner bottom bed area with more intrinsic models like 3-D and CFD would confirm whether the capture efficiencies are as good as the 1-D model predicts.
50
Acknowledgements 0
0 0
25
50
75
100
Flue gas load [%] Fig. 9. Development of the reactor average temperatures during different loads.
An interesting observation is that in the zero load situation fluidizing the carbonator with ambient air and running the calciner as an oxy-combustion circulating fluidized bed unit is also possible. Minimal heat was extracted from the external heat exchanger and solid train surfaces and the rest can be extracted in backpasses of the reactors. If the power plant is abruptly shut down, the calcium loop can provide back-up power for some period. To summarize the different load scenario analysis, flexibility can be achieved using solid and flue gas recirculation and clever heat transfer design. The flue gas flow out of the calciner in the zero load case drops to 60% of the full load flow which will certainly have effect on the backpass heat exchanger performance. The thermal efficiency in the partial loads will not be as good as predicted by the simplified approach used in this study. Also the
The research leading to these results has received funding from the Carbon Capture and Storage Programme (2011–2015), financed by the Finnish Funding Agency for Technology and Innovation (Tekes) and coordinated by the Finnish Cluster for Energy and Environment (CLEEN Ltd.). References [1] Metz B, Davidson O, de Coninck H, Loos M, Meyer L. Special report on carbon dioxide capture and storage. In: Intergovernmental panel on climate change. Cambridge, (UK): Cambridge University Press; 2005. [2] Sánchez-Biezma A, Ballesteros JC, Diaz L, de Zárraga E, Álvarez FJ, López J, et al. Postcombustion CO2 capture with CaO. Status of the technology and next steps towards large scale demonstration. Energy Proc 2011;4:852–9. [3] Galloy A, Ströhle J, Epple B. Design and operation of a 1 MWth carbonate and chemical looping CCS test rig. VGB Power Tech 2011;91:64–8. [4] Hawthorne C, Dieter H, Bidwe A, Schuster A, Scheffknecht G, Unterberger S, et al. CO2 capture with CaO in a 200 kWth dual fluidized bed pilot plant. Energy Proc 2011;4:441–8. [5] Shimizu T, Hirama T, Hosoda H, Kitano K, Inagaki M, Tejima K. A twin fluid-bed reactor for removal of CO2 from combustion processes. Chem Eng Res Des 1999;77:62–8.
J. Ylätalo et al. / Fuel 115 (2014) 329–337 [6] Dean CC, Blamey J, Florin NH, Al-Jeboori MJ, Fennell PS. The calcium looping cycle for CO2 capture from power generation, cement manufacture and hydrogen production. Chem Eng Res Des 2011;89:836–55. [7] Martínes A, Lara Y, Lisbona P, Romeo LM. Energy penalty reduction in the calcium looping cycle. Int J Greenshouse Gas Control 2012;7:4–81. [8] Li Y, Liu C, Sun R, Liu H, Wu S, Lu C. Sequential SO2/CO2 capture of calciumbased solid waste from the paper industry in the calcium looping process. Ind. Eng. Chem. Res 2012;51:16042–8. [9] Rodríguez N, Murillo R, Abanades JC. CO2 capture from cement plants using oxyfired precalcination and/or calcium looping. Environ Sci Techonol 2012;46:2460–6. [10] Florin N, Fennell P. Synthetic CaO-based sorbent for CO2 capture. Energy Proc 2011;4:830–8. GHGT-10. [11] Martínez I, Murillo R, Grasa G, Rodríguez N, Abanades JC. Conceptual design of a three fluidised beds combustion system capturing CO2 with CaO. Int J Greenhouse Gas Control 2011;5:498–504. [12] Nikolopoulos A, Nikolopoulos N, Charitos A, Grammelis P, Kakaras E, Bidwe AR, et al. High-resolution 3-D full-loop simulation of a CFB carbonator cold model. Chem Eng Sci 2013;90:137–50. [13] Vorrias I, Atsonios K, Nikolopoulos A, Nikolopoulos N, Grammelis P, Kakaras E. Calcium looping for CO2 capture from a lignite fired power plant. Fuel 2013, In press http://dx.doi.org/10.1016/j.fuel.2012.12.087. [14] Hawthorne C, Trossmann M, Cifre PG, Schuster A, Scheffknecht G. Simulation of the carbonate looping power cycle. Energy Procedia Greenhouse Gas Control Technologies 9. In: Proceedings of the 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), 16–20 November 2008, Washington, DC, USA, 2009;1:1387–94. [15] Abanades JC, Anthony EJ, Wang J, Oakey JE. Fluidized bed combustion systems integrating CO2 capture with CaO. Environ Sci Technol 2006;39:2861–6. [16] Romano M. Coal-fired power plant with calcium oxide carbonation for postcombustion CO2 capture. Energy Proc 2009;1:1099–106.
337
[17] Diego ME, Arias B, Abanades JC. Modeling the solids circulation rates and solids inventories of and interconnected circulating fluidized bed reactor system for CO2 capture by calcium looping. Chem Eng J 2012;198–199:228–35. [18] Martínez I, Murillo R, Grasa G, Abanades JC. Integration of a Ca-Looping system for CO2 capture in an existing power plant. Energy Proc 2011;4:1699–706. [19] Ylätalo J, Ritvanen J, Arias B, Tynjälä T, Hyppänen T. 1-Dimensional modeling and simulation of the calcium looping process. Int J Greenhouse Gas Control 2012;9:130–5. [20] Johnsson F, Leckner B. Vertical distribution of solids in a CFB-furnace, In: Heinschel K, editor, Proceedings of the 13th International Conference on Fluidized Bed combustion Orlando, FL, USA. May 07–10, 1995. p. 671–679. [21] Fang F, Li Z, Cai N. Experiment and modeling of CO2 capture from flue gases at high temperature in a fluidized bed reactor with Ca-based sorbents. Energy Fuels 2009;23:207–16. [22] Myöhänen K, Hyppänen T. A three-dimensional model frame for modelling combustion and gasification in circulating fluidized bed furnaces. Int J Chem Reactor Eng 2011;9. Article A25. [23] Alonso M, Rodríguez N, Gonzáles B, Grasa G, Murillo R, Abanades JC. Carbon dioxide capture from combustion flue gases with a calcium oxide chemical loop. Experimental results and process development. Int J Greenhouse Gas Control 2010;4:167–73. [24] Ylätalo J, Parkkinen J, Ritvanen J, Tynjälä T, Hyppänen T. Modeling of the oxycombustion calciner in the post-combustion calcium looping process. Fuel 2012. In press. http://dx.doi.org/10.1016/j.fuel.2012.11.041. [25] De-Souza Santos M. Solid fuels combustion and gasification. CRC Press; 2010. ISBN-9781420047493. [26] González B, Alonso M, Abanades JC. Sorbent attrition in a carbonation/ calcination pilot plant for capturing CO2 from flue gases. Fuel 2010;89:2918–24.