Simulation of a Calcium Looping CO2 Capture Process for Pressurized

Fluidized Bed Combustion

Benoit Duhoux†, Robert T. Symonds‡∗, Robin Hughes‡, Poupak Mehrani†, Edward J. Anthony
§
, Arturo Macchi†

† Department of Chemical and Biological Engineering, Centre for Catalysis Research and
Innovation, University of Ottawa, 161 Louis Pasteur Street, Ottawa, Ontario K1N 6N5 Canada
‡ Natural Resources Canada, CanmetENERGY, 1 Haanel Drive, Ottawa, Ontario K1A 1M1
Canada
§
School of Applied Science, Cranfield University, Cranfield, Bedfordshire MK43 0AL, United
Kingdom

ABSTRACT

The Canadian regulations on carbon dioxide emissions from power plants aim to lower the
emissions from coal-fired units down to those of natural gas combined cycle (NGCC) units.
Since coal is significantly more carbon intensive than natural gas, coal-fired plants must
operate at higher net efficiencies and implement carbon capture to meet the new regulations.
Calcium looping (CaL) is a promising post-combustion carbon capture (PCC) technology that,
unlike other capture processes, generates additional power. By capturing carbon dioxide at
elevated temperatures, the energy penalty that carbon capture technologies inherently impose
on power plant efficiencies is significantly reduced. In this work, the CO2 capture performance
of a calcium-based sorbent is determined via thermogravimetric analysis under relatively high
carbonation and low calcination temperatures. The results are used in an aspenONE™
simulation of a CaL process applied to a pressurized fluidized bed combustion (PFBC) system
at thermodynamic equilibrium. Combustion of both natural gas and coal are considered for
sorbent calcination in the CaL process. A sensitivity analysis on several process parameters,
including sorbent feed rate and carbonator operating pressure, is explored. The energy penalty
associated with the capture process ranges from 6.8 to 11.8 percentage points depending on
fuel selection and operating conditions. The use of natural gas results in lower energy penalties
and solids circulation rates, while operating the carbonator at 202 kPa(a) results in the lowest
penalties and drops the solids circulations rates to below 1000 kg/s.
KEYWORDS: Carbon capture, calcium looping, pressurized FBC, process simulation

∗Corresponding author: Robert Symonds, Tel: +1 613 996-5153, Fax: +1 613 992-9335
E-mail address: [email protected]
INTRODUCTION

Fossil fuels represent the major energy source for electricity production in the world. In 2015,
approximatively 66% of the world production relied on fossil fuel, with coal alone accounting
for up to 39% of the global power output.[1] Whilst energy production predictions foresee a
shift towards renewable sources of energy (e.g., solar, hydro, wind, etc.), fossil fuels, and coal
in particular, will still constitute an important source of energy.[1] Natural gas represents a
cleaner source of energy as it is less carbon intensive,[2] whereas coal contains more sulphur
and generates ash, along with a wide range of micro-pollutants (e.g., halides, mercury). The
CO2 intensity of a process, defined as the amount of CO2 emitted per unit of net power
produced, depends mainly on the fossil fuel burned in the power plant. Average values range
from about 400 kg/MWeh for natural gas units to 900 kg/MWeh for coal cases.[2] Present
regulations in Canada require that carbon emissions from existing coal power plants reach
natural gas-fired levels by 2030 or the end of their useful life (defined as between 45 and 50
years after commissioning date), whichever comes first. New coal-fired power plants must
comply upon start-up.[3,4] The important difference in current carbon emissions from these two
types of power plants will force the implementation of carbon capture and storage (CCS) in
coal-fired power plants, in addition to efficiency improvements.

Pressurized fluidized bed combustion (PFBC) is a particular implementation of fluidized bed

combustion where the boiler is usually operated between 1 and 1.5 MPa.[5] By operating at
elevated pressures, additional energy recovery via latent heat and hot flue gas expansion can
be realized, in addition to the steam cycle used in all power plants. This combined cycle results
in efficiencies above 40% on a higher heating value (HHV) basis, reducing the carbon intensity
of the process.[5,6] Calcium looping (CaL) is a post-combustion carbon capture technology that
uses the reversible carbonation reaction to capture CO2. The process circulates calcium-based
sorbents between two reactors: the carbonator where CO2 reacts with calcium oxide (Equation
(1)); and the calciner where the combustion of a fuel generates the heat necessary to calcine
calcium carbonate (Equation (2)). The process is illustrated in Figure 1. Due to sorbent
sintering and/or reaction with other species (e.g., SO2), the conversion of calcium oxide in the
carbonator decreases as the sorbent is cycled in the process.[7,8] To maintain the conversion at
a desired level, a make-up of fresh limestone and a purge to remove spent sorbent are
implemented. The calciner operates in oxy-combustion mode to avoid the dilution of CO2 with
nitrogen from air and to reduce the energy required for the compression of the calciner outlet
gas.
 ( ) + ( ) → ( ) ∆ ° = −173.3 / (1)

( ) → ( ) + ( ) ∆ ° = 169.5 / (2)

Figure 1: Calcium looping (CaL) block flow diagram.

This work considers the simulation of a CaL process for post-combustion CO2 capture from a
PFBC power plant. Mass and energy balances are performed to investigate the effect of
operating parameters on stream characteristics via a series of sensitivity analyses. Estimations
for the energy penalty of CaL on the base power plant are also provided as an additional
comparison criterion. In this investigation, the gross efficiency associated with power
production from the carbon capture process is based on previous work on a similar system.[9]
The capture process simulations are based on the cyclic CO2 sorption of a limestone that was
investigated in a series of thermogravimetric analyzer (TGA) experiments. The limestone was
subjected to multiple cycles of relatively high-temperature carbonation and low-temperature
calcination. These particular conditions were selected in order to minimize the temperature
difference between the two reactors in the CaL process and reduce the fuel consumption in the
calciner. The use of both natural gas and coal in the calciner was compared by using different
gas atmospheres during the calcination step.

MATERIAL AND METHODS

Experimental Methods
The tests were conducted using a naturally occurring Canadian limestone (Cadomin), sieved to
a diameter range of between 300 and 425 µm, which is an appropriate size range for circulating
fluidized bed operation. The composition of the sorbent, obtained from X-ray fluorescence
(XRF), is provided in Table 1.

Table 1: Composition of Cadomin limestone as determined via XRF (300 – 425 µm in

diameter)

Mass fraction
Component
(wt%)

SiO2 1.44

Al2O3 0.3

Fe2O3 0.35

TiO2 <0.03

P2O5 <0.03

CaO 52.27

MgO 1.54

SO3 <0.10

Na2O <0.20

K2O 0.09

Ba <250 ppm

Sr 258 ppm

V <50 ppm

Ni <50 ppm

Mn 80 ppm

Cr < 50 ppm
 Cu <30 ppm

Zn <30 ppm

LOF* 43.96

*Loss on fusion

The sorbent was tested using a Cahn 1000 Electrobalance thermogravimetric analyzer, using
sample sizes of approximately 20 mg. This particular sample size was chosen in order to lower
potential variance in chemical composition between runs, as individual particles display
different CaO/MgO ratios. The total gas flowrate in the furnace was maintained at 100
NmL/min using mass flow controllers. The gas compositions were switched between
carbonation and calcination cycles as shown in Table 2. Based on previous testing, where CaO
carbonation profiles were compared at multiple flowrates ranging from 50 to 200 NmL/min, it
has been shown that the selected operating conditions limit gas diffusion resistance from the
bulk gas to the particle surface.[10,11] In addition, Perejon et al.[12] performed similar testing
using a gas flow rate of 50 mL/min and noted that gas diffusion resistance through the sample
only becomes relevant for sample masses above 40 mg for their TGA setup.

Table 2: TGA test conditions

Carbonation Calciner

Gas Gas
Temperature Pressure Temperature Pressure
Case composition composition
(°C) (kPa(a)) (°C) (kPa(a))
(vol.%) (vol.%)

Base CO2 -16.7%

CO2 - 63.3%
Case 730 N2 - 76.4% 101 892 101
H2O - 36.7%
NG H2O - 6.9%

CO2 -16.7%
Press. CO2 - 64.5%
778 N2 - 76.4% 202 893 101
NG H2O - 35.5%
H2O - 6.9%
 Base CO2 -16.7%
CO2 -84.8%
Case 730 N2 - 76.4% 101 912 101
H2O - 15.2%
Coal H2O - 6.9%

CO2 -16.7%
Press. CO2 - 85.3%
778 N2 - 76.4% 202 912 101
Coal H2O - 14.7%
H2O - 6.9%

The gas compositions and temperatures were selected using preliminary process simulations to
identify conditions where the carbonator and calciner would operate with a minimal
temperature difference. The tests simulate both natural gas and coal flue gas generated during
oxy-fuel calcination. Tests were performed at atmospheric pressure and 202 kPa(a) for both
simulated fuels. At atmospheric pressure, an initial 40-min calcination was performed,
followed by subsequent 5-min carbonations and 5-min calcinations. When carbonating at 202
kPa(a), the carbonation duration was increased to 10 min to ensure the end of the kinetic-
controlled stage of carbonation is reached. At the end of the carbonation period, the system
pressure was slowly dropped to 101 kPa(a) using an automated pressure control valve which
ensured the target gas composition was maintained. After calcination at atmospheric pressure,
the reactive gas was switched to an inert and the pressure was increased to 202 kPa(a) via the
same automated pressure control valve. In all cases, 20 cycles were performed on each sample.

A typical carbonation/calcination cycle is illustrated in Figure 2. Once the carbonation

temperature is reached, the gas atmosphere is switched and the weight of the sample starts
increasing after a short delay due to the volume of the gas lines. The carbonation is divided
into two steps: a fast reaction kinetic-driven stage followed by a slower diffusion-limited stage.
Appearance of the second step results from the formation of a carbonate layer at the surface of
the particles through which CO2 must diffuse before accessing unreacted CaO.[13] After the
carbonation period, the temperature is increased and the gas atmosphere switched to calcination
conditions. Due to the higher CO2 partial pressure, the sample weight increases (via
carbonation) until a sufficiently high temperature is reached for calcination to occur; the sample
temperature is raised to the calcination temperature as quickly as possible to limit its impact.
The level of recarbonation in Figure 2 (first cycle) is not observed in the subsequent cycles
and, therefore, does not significantly affect the activity of the sorbent, which ultimately reaches
its residual value. It should be noted that this extra carbonation is unavoidable when calcining
under high CO2 environments and has been noted by previous authors when operating at
temperatures above 900oC and 70 vol.% CO2.[12,14] Some level of recarbonation can be
expected in a real system during transfer of the sorbent to the calciner (using CO2) and has been
observed experimentally.[15] The sample is fully calcined before the end of the calcination
period, after which the temperature is reduced under pure nitrogen before initiating the next
cycle.

6 1000
Data
Temperature 950
5
Weight change (mg)

900

Temperature (oC)
4
Carbonation Calcination
850
Region Region
3
800
2
Diffusion 750
Controlled
1
700
Kinetic
Controlled
0 650
95 100 105 110 115 120
Time (min)

Figure 2: Sample weight and temperature vs. time for the first cycle of the simulated coal
base case conditions – Carbonation in 76.4% N2, 16.7% CO2 and 6.9% H2O – Calcination
in 84.8% CO2 and 15.2% H2O. All concentrations in vol.%.
It should be noted that the CaO conversion used in this work only takes into account the first
phase of the carbonation. This choice was made to reflect the shorter residence times of the
particles in the circulating fluidized bed that the authors expect to be used for this CO2 capture
process. Consequently, the expected average conversions are lower than what is presented in
the available literature, especially from TGA data.[16–18]

Computational Methods

The PFBC was modeled using information from the Karita P800 boiler in Japan that was
designed to produce 360 MWe with a net efficiency of 41.8% HHV.[6,19] The power output is
split between a steam turbine (290 MWe) and a gas turbine (70 MWe). The key specifications
of this unit are summarized in Table 3.

Table 3: Main specifications of Karita P800 from Koike et al. [19]

Boiler type Pressurized bubbling fluidized bed

Total output 360 MWe

Gas turbine output 70 MWe

Steam turbine output 290 MWe

Net efficiency 41.8% HHV

Gross efficiency 42.8% HHV

Steam conditions (supercritical)

Pressure 24,100 kPa(a)

Temperature 566/593°C (Main/reheat)

Gas conditions

Pressure Approx. 1300 kPa

Temperature Approx. 850oC

De-SOx system In-situ with limestone

 De-NOx system PFBC combustion + SCRa

Dust collector 2-stage cyclones + ESPb

a
selective catalytic reduction; belectrostatic precipitator

The fuel fed to the boiler contains a lime-based sorbent for in-situ desulphurization.[6] The
combustion takes places at 1300 kPa(a) and 850°C. A two-stage cyclone is used to reduce the
particle matter concentration before sending the flue gas to the gas turbine. After the turbine,
the gas is treated for NOx using a selective catalytic reduction process (SCR) and passes
through the steam cycle economizer used for boiler feed water pre-heat. A last particle removal
step is carried out with an electrostatic precipitator (ESP) before sending the flue gas to the
stack.

The Karita PFBC is simulated using Blair Athol coal, with proximate and ultimate analyses
provided in Table 4. The total coal feed rate was adjusted to match the desired power output of
360 MWe. Note that the total coal feed rate to the PFBC unit is directly calculated from the fuel
HHV and the net efficiency of the PFBC via the following equation:

̇ = (3)
∙

where ̇ is the total coal feed rate, is the desired power output, is the higher heating
value of the coal, and is the net efficiency of the PFBC unit. The sorbent feed rate is based
on results from the operation of the Tidd PFBC demonstration project where a calcium-to-
sulphur (Ca/S) molar ratio of 1.5 was sufficient to achieve 95% sulphur capture.[20]
Table 4: Blair Athol coal analysis from Huleatt [21]

Proximate analysis Ash analysis (wt.%)

Moisture wt.% 7.5 SiO2 61.3

Volatile matter wt.% 27.2 Al2O3 30.2

Fixed carbon wt.% 57.3 Fe2O3 4.2

Ash wt.% 8 CaO 0.5

Total wt.% 100 MgO 0.5

TiO2 1.6

Ultimate analysis Na2O 0.2

Moisture wt.% 7.5 K2O 0.3

Carbon wt.% 68.6 Mn3O4 0.1

Hydrogen wt.% 3.7 P2O5 0.2

Nitrogen wt.% 1.4 SO3 0.2

Chlorine wt.% 0

Sulphur wt.% 0.3

Ash wt.% 8

Oxygen wt.% 10.5

Total wt.% 100

Calorific analysis (HHV, air dried) MJ/kg 27.28

Combustion air is compressed from atmospheric pressure to 1300 kPa(a), with an assumed
composition of 79 vol.% nitrogen and 21 vol.% oxygen. Following the quality guidelines for
energy system studies suggested by NETL,[22] the stage pressure ratio is limited to 2.5, meaning
that a three-stage compressor is required. Inter-stage cooling is implemented so that the
discharge air temperature from each stage does not exceed 167°C.[23] A 2% pressure drop based
on the inlet pressure is assumed for each inter-stage cooler.[22] The boiler is simulated using
the Peng-Robinson equation of state and a Gibbs reactor with a combustion efficiency of 99.5%
to match the performance of the Tidd unit.[20] The isentropic efficiency of the turbine is
calculated by reproducing the Karita unit performance. The efficiency is adjusted until the
power output from the gas turbine is 70 MWe, resulting in an isentropic efficiency of 88%. For
the present work, the efficiency calculated above is adopted for all gas turbines in the
simulation. The discharge pressure is set to 121.3 kPa(a), to account for an assumed pressure
drop of 20 kPa in the calcium looping carbonator. The mechanical efficiency of the turbine is
set at 98.5%.[22] The output from the gas turbine is then fed to the CaL capture cycle as
illustrated in Figure 3.

Process simulations were performed using aspenONETM, with the base cases including the use
of both coal and natural gas as fuels for the calciner. Between these two simulations, all the
operating conditions were identical, with the exception of the fuel due to the presence of
sulphur and ash for the coal cases. The sulphur causes a slight increase in the CaO circulation
rate from the irreversible formation CaSO4. Note that the influence of ash on sorbent
conversion is not considered here and is assumed negligible based on low measured cyclic
conversions and relatively high sorbent make-up rates. In the simulation, the ash increases the
amount of inert solids in the process, leading to higher fuel consumption. Base case simulation
parameters for natural gas and coal cases are presented in Table 5 where and are defined
as the molar flow of CaCO3 in the make-up stream and the molar flowrate of CO2 in the flue
gas entering the carbonator, respectively. The reactors operate at atmospheric pressure and at
temperatures corresponding to the experiments conducted and presented in the experimental
section. Conversion of the CaO in the carbonator is tied to the solids circulation rate, the make-
up rate, and the sorbent properties.
Figure 3: PFBC-CaL process flow diagram

Table 5: Base case simulation parameters

Reactor Carbon
Pressure Make- Intensity
Carbonator Calciner
Parameter/Fuel Drop up Ratio
Temperature Temperature (kg
(F0/FCO2)
CO2/MWeh)

NG 730°C 892°C 20 kPa 0.05 420

Coal 730°C 912°C 20 kPa 0.12 420

Due to the sorbent deactivation during particle cycling, an estimation of the particle residence
time in the process is required. In addition, the following equation is utilized to predict the
particle conversion, , over a large number of process cycles, :[24]

= + (4)

where is a decay parameter and is the residual conversion at high cycle numbers. The
average conversion of all the solids in the carbonator, , is a weighted average over the
cycle numbers:[25]

=∑ (5)
 =( )
(6)

where represents the fraction of particles of age , is the molar flow rate of CaO entering
the carbonator. The sum in Equation (5) is an infinite sum. During this study, enough cycles
have been considered to account for over 95% of the solids (∑ ≥ 0.95). The make-up rate
of limestone (Equation (7)) and the circulation of calcium oxide (Equation (8)) are both defined
as ratios based on flue gas CO2 entering the CaL process:

(7)

(8)

All cases are simulated under conditions that meet the Canadian carbon emissions regulations
currently in place.[3] No hydrodynamic modeling of the carbonator has been included, instead
the carbonator is assumed to be perfectly mixed and the partial pressure of the CO2 leaving the
reactor is at equilibrium conditions. Therefore, the carbonator CO2 capture efficiency was
dependant on the carbonator temperature and varied based on operating pressure. The amount
of circulating CaO and the fuel feed rate are calculated so that the calciner does not produce
additional heat (i.e., only enough heat for calcination) and to meet the carbon intensity
specification. The production of oxygen (99 vol.%) is not simulated and the air separation unit
(ASU) energy consumption is calculated from a specific energy consumption of 231
kWh/tonne O2.[26] The oxygen is fed in excess to maintain 2.7 vol.% O2 (dry basis) in the gas
outlet of the calciner.[22]

Calciner flue gas recycling is implemented to maintain fluidization conditions and reduce the
oxygen concentration in the fluidization gas. While the process simulation does not include a
hydrodynamic model or sizing of the calciner, the recycle ratio is set to achieve 30 vol.% O2
concentration at the reactor windbox, in alignment with the base case of the US Department of
Energy (DOE) experimental work.[26] In all cases, the calcination conversion was fixed at
100%. The recycled flue gas stream is cooled to limit the blower inlet temperature to 190°C.
The blower compensates for the calciner pressure drop and increases the gas stream pressure
by 20 kPa. The energy required to condition and compress the concentrated CO2 stream is
calculated using a 5-stage compression process with inter-stage cooling. The condensate
collected after each cooling step is separated before sending the gas phase to the next
compression stage. The CO2 stream is compressed in this manner until condensation occurs at
20°C, after which the CO2 is pumped to a final pressure of 15000 kPa(a). A dedicated sub-
critical steam cycle is used to recover energy from the capture process and produce additional
power. The gross efficiency of that cycle was calculated to be 35% (HHV) in previous work,[9]
and has been employed here.
RESULTS

Limestone Conversion

As discussed in the experimental method section above, only sorbent performance during the
initial rapid carbonation stage is important, since marginal improvements in carbonation
conversion are realized after this point. The end of this stage is approximated by the intersection
of two linear regressions for each of the carbonation stages. The results presented here illustrate
the decay of the sorbent conversion with repeated process cycles. The emphasis is on
conversion at high cycle numbers, since high make-up ratios are not desirable from an
economic perspective unless the spent lime can be used for cement applications for example
which is outside of the remit for this paper. Results from the experimental tests simulating the
use of natural gas for calcination are presented in Figure 4. It was observed that operating the
carbonation at a higher pressure results in higher conversion, due largely to the higher CO2
partial pressure. The improvement over the results at atmospheric pressure tend to be reduced
at higher cycle numbers, although the predicted values for residual conversions are still higher
under pressurized conditions (Table 6). This is an important result since conversion predictions
at high cycle numbers greatly affect process performance.

The experiments simulating coal combustion for sorbent calcination show a similar behaviour
(Figure 5). The higher CO2 partial pressures result in higher conversions at all cycle numbers
(Table 6). Comparing coal and natural gas conditions, we can see that natural gas calcination
conditions are slightly more favourable in terms of CaO conversion, likely due to a combination
of lower operating temperature and the presence of steam. While the individual effects of these
two parameter are not tested here, they have both been tested independently elsewhere and
shown to increase CaO conversion.[16, 27]
Although, the natural gas conditions result in
marginally better CaO conversions, both sets of calcination conditions give relatively low
cyclic conversions, even over the first few cycles. As noted above, the recorded CaO
conversions are taken at the end of the kinetic-controlled stage and partially explain the reduced
cyclic carrying capacity. In addition, calcination takes place under high CO2 partial pressures
and elevated temperatures, which are known to increase particle sintering, limiting particle
porosity, and dramatically decreasing the rate of carbonation. Perejon et al. [12] and Ortiz et al.
[14]
have noted that the traditional equations for calculating CaO conversion do not reflect the
effect of calcination under high CO2 partial pressures, as would be the case in the CaL
technologies, and that the solid-state diffusion controlled carbonation is “greatly enhanced”
when CaO is regenerated under high CO2 partial pressures as compared to calcination under
low CO2 concentration. Their results showed CaO conversions, at the end of the reaction-
controlled stage, of ~4% after 20 cycles and residual conversions of <3% under similar
conditions as the present work. In addition, these low residual activities are in-line with
carbonation performance observed during pilot-scale trials with the same sorbent.[15]

0.20 Natural gas - 101 kPa

0.18 Natural gas - 101 kPa - Model

Natural gas - 202 kPa
Conversion (mass fraction)

0.16
Natural gas - 202 kPa - Model
0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00
0 5 10 15 20
Cycle Number

Figure 4: Sorbent conversion vs. process cycle number – Simulated natural gas flue gas
conditions.

0.20

0.18 Coal - 101 kPa

Coal - 101 kPa - Model
Conversion (mass fraction)

0.16 Coal - 202 kPa

0.14 Coal - 202 kPa - Model

0.12

0.10

0.08

0.06

0.04

0.02

0.00
0 2 4 6 8 10 12 14 16 18 20
Cycle Number
Figure 5: Sorbent conversion vs. process cycle number – Simulated coal flue gas
conditions.
The conversion decay model presented in the experimental method section was fitted to all four
sets of data, with the resulting model parameters presented in Table 6. Particular attention has
been given to how well the model fits the higher cycle number data points (>10 cycles) due to
the importance of high cycle number on system performance. While complete data sets
obtained at atmospheric pressure were used during model fitting, the two pressurized data sets
were fitted based only on the last 10 to 15 cycle data points. This decision was made based on
the significant variability in sorbent conversion, which is thought to be due to greater water
throughput. With increased water injection rates to the system, the likelihood of incomplete
vapourization in the feed line increases, allowing liquid to enter the TGA. The authors believe
this rapid vapourization caused a perturbation in the gas flow around the sample, causing a
significant amount of noise in the recorded sample weight. Although the decay model provides
a poor fit for the first few cycles, this does not affect the subsequent modeling results since the
average sorbent conversion of all the solids in the carbonator is effectively equal to the residual
conversion.

Table 6: Sorbent conversion model parameters

Natural gas cases Coal cases

P = 101 P = 202 P = 101 P = 202

kPa kPa kPa kPa

0.026 0.031 0.017 0.020

k 25.012 5.832 9.792 2.994

Process Simulations

The results of the base case simulations for atmospheric pressure natural gas calcination are
shown in Table 7. Process performance metrics are presented for two distinct scenarios: (1)
where the current regulation is targeted (i.e., the power required by the CO2 processing unit
(CPU) is left out of the energy penalty); and (2) a more stringent case where the compression
of CO2 is accounted for in the CO2 intensity calculations. The focus of the discussion will be
on scenario (1), but results from both are presented.
Table 7: Base case simulations – Natural gas – PFBC-CaL process performance summary
with and without the CPU included in target CO2 emissions intensity

With CPU Without CPU

Net power (MWe) 360.0 360.0
PFBC
Efficiency (% HHV) 41.8 41.8
Thermal input (MWth) 490.8 461.3
CaL assumed gross efficiency
35.0 35.0
(% HHV)
CaL gross power (MWe) 171.8 161.5
CaL Process

Auxiliaries (MWe) 11.8 11.1

CPU (MWe) 21.2 20.0
ASU (MWe) 32.1 30.2
Net power (MWe) 106.7 100.2
Efficiency (% HHV) 21.7 21.7
Solids flow rate (kg/s) 1212.1 1132.4
CaO conversion (%) 2.6 2.6
Global net power (MWe) 466.7 460.2
Global Process

Global efficiency (% HHV) 34.5 34.8

Energy penalty (% HHV) 7.3 7.0
CO2 intensity (kg CO2/MWeh) 420.0 420.0

The simulated CaL capture process produces an additional 161.5 MWe of power (gross) of
which 26% is consumed to produce oxygen and drive the process auxiliaries. The overall
efficiency, grouping both the capture process and the power plant, was calculated to be 34.8%.
This represents an energy penalty of 7.0 percentage points over the base power plant efficiency
(41.8%). Due to the low sorbent make-up ratio selected for the base case, the sorbent particles
are cycled a large number of times and the resulting average conversion reaches the residual
conversion of 2.6%, as determined experimentally. This low cyclic conversion results in a
relatively high solid circulation rate (1132 kg/s). In terms of loading, a 2.6% cyclic conversion
equates to a capture rate of 20.4 g CO2/kg sorbent. Although it does not compare favourably to
amine absorption (17.6-40.5 g CO2/kg solvent)[28] or zeolite adsorption (37.4-48.4),[29] the
carbonation of the limestone in the simulation only accounts for the fast carbonation phase, as
explained in the experimental methods section.

The use of coal to provide thermal energy for calcination was also investigated (Table 8). In
this case, the sorbent make-up ratio was adjusted to control the circulating ash in the process.
The ash leaving the process in the purge stream causes higher purge/make-up flows to limit
ash accumulation in the system. In the base case, ash represents 15 wt.% of the circulating
solids at a make-up ratio of 0.12. In addition to ash impeding access to the active CaO sites,
the presence of sulphur also results in the formation of CaSO4. Together, the ash and CaSO4
contribute to the overall solids circulation. Compared to the natural gas case, the solids
circulation rate is significantly increased due to the presence of these inert solids and the lower
average sorbent conversion.

Using coal also causes a larger disparity between the current CO2 emissions intensity regulation
and the more stringent cases. Due to a larger volume of CO2, the power used to run the CPU
increases and the difference in energy penalty between the two capture targets increases from
0.3 (natural gas case) to 0.9 percentage points. The absolute energy penalty of the capture
process is also higher for coal via increased oxygen consumption and total CPU demands.
Table 8: Base case simulations – Coal – PFBC-CaL process performance summary with
and without the CPU included in target CO2 emissions intensity

With CPU Without CPU

Net power (MWe) 360.0 360.0
PFBC
Efficiency (% HHV) 41.8 41.8
Thermal input (MWth) 585.9 513.1
CaL Assumed gross efficiency
35.0 35.0
(% HHV)
CaL Gross power (MWe) 205.1 179.6
Auxiliaries (MWe) 15.8 13.8
CaL Process

CPU (MWe) 45.1 39.5

ASU (MWe) 53.7 47.0
Net power (MWe) 90.5 79.3
Efficiency (% HHV) 15.4 15.4
Solids flow rate (kg/s) 1983.5 1737.2
CaO conversion (%) 2.1 2.1
F0/FCO2 0.14 0.12

Global net power (MWe) 450.5 439.3

Global Process

Global efficiency (% HHV) 31.1 32.0

Energy penalty (% HHV) 10.7 9.8
CO2 intensity (kg CO2/MWeh) 420.0 420.0

A sensitivity analysis on the sorbent make-up ratio, which significantly influences the
economics of the process, [30] was performed to evaluate its impact on process performance.
The make-up rate, , was varied from 0.01 to 0.1, while adjusting the overall solids circulation
rate to meet the specified carbon emissions intensities. The results of this analysis for the
natural gas calcination cases are provided in Table 9.

Table 9: Sensitivity analysis – Sorbent make-up ratio – Natural gas cases with and
without the CPU included in target CO2 emissions intensity

With CPU Without CPU

 F0 = F0 = F0 = F0 = F0 = F0 = F0 = F0 = F0 = F0 =
0.01 0.025 0.05 0.075 0.1 0.01 0.025 0.05 0.075 0.1
Solids flow
1261.2 1241.8 1212.1 1184.3 1157.6 1186.4 1165.1 1132.4 1101.4 1071.5
rate (kg/s)
CaO
2.57 2.60 2.64 2.68 2.72 2.58 2.61 2.65 2.69 2.73
conversion (%)
Global process
net power 463.7 464.8 466.7 468.6 470.6 457.7 458.6 460.2 461.9 463.6
(MWe)
Global process
efficiency (% 34.7 34.7 34.5 34.4 34.2 35.0 34.9 34.8 34.7 34.5
HHV)
Energy penalty
7.1 7.1 7.3 7.4 7.6 6.8 6.9 7.0 7.1 7.3
(% HHV)
CO2 intensity
(kg 420.0 420.0 420.0 420.0 420.0 420.0 420.0 420.0 420.0 420.0
CO2/MWeh)

Higher sorbent make-up ratios result in higher sorbent conversion and lower solids circulation
rates, as a large portion of the particles are cycled fewer times. However, fuel and oxygen
consumptions increase in order to heat the make-up stream to the calciner temperature and
provide the necessary heat for calcination of the fresh limestone. Consequently, additional
power is generated from the CaL capture process and the proportion of low-efficiency power
increases in the overall system. This ultimately increases the energy penalty at higher sorbent
make-up ratios.

For the coal cases, the sensitivity analysis is presented in terms of ash content in the circulating
solids (Table 10). To reduce the ash content of the solid stream, a higher purge rate and, hence,
a higher make-up ratio are required. The same effect on sorbent conversion and solid
circulation rate is then observed. In this case, however, due to the range in which the make-up
ratio is modified, a point is reached where the energy penalty is minimized (~15 wt.% ash
content). At lower ash contents, the energy required to heat and calcine the fresh limestone
becomes prohibitive. At higher ash contents, the amount of inert solids circulating in the
capture process significantly increases the fuel and oxygen demands via sensible heating
demands, resulting in higher penalties.

Operating the carbonator at elevated pressures has been shown to increase the sorbent
utilization potential. Considering the challenges with the transfer of solids between reactors
operating at different pressures, a case with the carbonator operating at 202 kPa(a) was
simulated. This is expected to be feasible since this minimal pressure difference should be
easily controlled in the transfer lines. A major benefit realized by operating at elevated
pressures is that the vitiated flue gas leaving the carbonator can be expanded for additional
power generation. Under these conditions, the power produced by the power plant gas turbine
is reduced due to the higher discharge pressure, but is more than made up for by the expander
downstream of the carbonator. The total power produced from the combination of two turbines
is higher than in the base case, as the carbonator acts as a reheat for the flue gas before the final
expansion. The cost of the additional turbine, and its ultimate impact on the cost of electricity,
should be addressed in future economic studies.
Table 10: Sensitivity analysis – Circulating solids ash content – Coal cases with and
without the CPU included to obtain the target CO2 emissions intensity

With CPU Without CPU

Ash = Ash = Ash = Ash = Ash = Ash = Ash = Ash =
5 wt.% 10 wt.% 15 wt.% 20 wt.% 5 wt.% 10 wt.% 15 wt.% 20 wt.%
Solids flow
1242.3 1726.7 1983.5 2181.9 1049.3 1503.4 1737.2 1912.6
rate (kg/s)
CaO
conversion 3.05 2.28 2.08 1.98 3.05 2.28 2.08 1.98
(%)
F0/FCO2 0.50 0.21 0.14 0.10 0.42 0.19 0.12 0.09
Global net
446.9 446.3 450.5 455.3 433.4 435.2 439.3 443.6
power (MWe)
Global
efficiency (% 30.0 31.0 31.1 31.1 31.1 31.9 32.0 31.9
HHV)
Energy
penalty (% 11.84 10.78 10.67 10.75 10.7 9.9 9.8 9.9
HHV)
CO2 intensity
(kg 420.0 420.0 420.0 420.0 420.0 420.0 420.0 420.0
CO2/MWeh)

For both the natural gas and coal cases, the effect of increasing the carbonator operating
pressure is beneficial on multiple fronts, as shown in Tables 11 and 12. Increasing the power
output from the turbines reduces the amount of CO2 requiring capture to reach the regulation
target. In addition, the fuel demand is significantly reduced, along with oxygen consumption
and power requirements for auxiliaries and CO2 processing. Sorbent conversions under
pressurized conditions are higher, which contributes to the reduction in solids circulation and
lowers the capture level requirement. In the natural gas case, pressurization lowers the
circulating rate; however, this is not the case for coal since the make-up ratio is reduced at
higher pressure to maintain the 15% ash content, causing an increase in the solids flow rate.
Table 11: Sensitivity analysis – Pressure – Natural gas cases with and without the CPU
included in the target CO2 emissions intensity

With CPU Without CPU

P = 101 P = 202
P = 101 kPa P = 202 kPa
kPa kPa
Thermal input (MWth) 490.8 368.9 461.3 347.3
CaL Assumed gross
35.0% 35.0% 35.0% 35.0%
efficiency (% HHV)
CaL Gross power
171.8 129.1 161.5 121.6
(MWe)
CaL process

Auxiliaries (MWe) 11.8 8.9 11.1 8.4

CPU (MWe) 21.2 19.1 20.0 18.0

ASU (MWe) 32.1 24.3 30.2 22.8

Net power (MWe) 106.7 76.9 100.2 72.3
Efficiency (% HHV) 21.7% 20.8 21.7% 20.8
Solids flow rate (kg/s) 1212.1 955.9 1132.4 890.2
CaO conversion 2.6% 3.5% 2.6% 3.6%
Global net power
466.7 457.2 460.2 452.8
(MWe)
Global efficiency (%
Global process

34.5% 37.2% 34.8% 37.5%

HHV)
Energy penalty (%
7.3% 7.0% 7.0 6.7%
HHV)
CO2 intensity (kg
420.0 420.0 420.0 420.0
CO2/MWeh)

Table 12: Sensitivity analysis – Pressure – Coal cases with and without the CPU included
in the target CO2 emissions intensity

With CPU Without CPU

P = 101 P = 202
P = 101 kPa P = 202 kPa
kPa kPa
 Thermal input (MWth) 585.9 482.4 513.1 426.7
CaL Assumed gross
35.0 35.0 35.0 35.0
efficiency (% HHV)
CaL Gross power
205.1 168.9 179.6 149.3
(MWe)
Auxiliaries (MWe) 15.8 13.0 13.8 11.5
CaL process

CPU (MWe) 45.1 39.0 39.5 34.5

ASU (MWe) 53.7 44.3 47.0 39.2
Net power (MWe) 90.5 72.5 79.3 64.2
Efficiency (% HHV) 15.4 15.0 15.4 15.0
Solids flow rate (kg/s) 1983.5 2063.6 1737.2 1825.0
CaO conversion (%) 2.08 2.01 2.08 2.01
F0/FCO2 0.14 0.11 0.12 0.10
Global net power
450.5 448.9 439.3 440.9
(MWe)
Global efficiency
Global process

31.1 33.4 32.0 34.2

(% HHV)
Energy penalty (%
10.7 10.3 9.8 9.5
HHV)
CO2 intensity (kg
420.0 420.0 420.0 420.0
CO2/MWeh)
CONCLUSIONS

Cadomin limestone performance for CaL-based CO2 capture was investigated in a TGA using
relatively high-temperature carbonation and low-temperature calcination. Considering only the
first fast phase of carbonation, the sorbent was exposed to conditions representing both natural
gas and coal combustion flue gases. In addition, sorbent performance at elevated pressure
carbonation conditions was investigated. The results showed residual sorbent conversions
ranging from 2 to 3%, with simulated coal flue gas achieving lower conversions than those of
natural gas.

A CaL-based CO2 capture process was modeled using process simulation software to assess
the impact of post-combustion CO2 capture, using both natural gas and coal, on the efficiency
of a PFBC power plant. Natural gas performance was superior in all cases due primarily to the
absence of ash and CaSO4, which led to higher solids circulation rates and energy penalties. A
sensitivity analysis on the sorbent make-up ratio was investigated. In the case of natural gas, a
reduction in solids circulation was observed, ultimately increasing the energy penalty via fresh
sorbent heating demands. On the other hand, coal cases showed a minimum energy penalty at
around 15 wt.% ash content. Increasing the operating pressure of the carbonator was beneficial
from both an energy penalty and solids circulation perspective, with the exception of the coal
case where maintaining the same ash content requires a lower sorbent make-up ratio; thus,
causing a higher circulation rate.

ACKNOWLEDGEMENT

This research was sponsored and funded by the Program for Energy Research and
Development (PERD) at Natural Resources Canada, Government of Canada.
REFERENCES

[1] International Energy Agency, "Key World Energy Statistics," IEA, 2017, accessed on
10 August 2018,
https://www.iea.org/publications/freepublications/publication/KeyWorld2017.pdf.

[2] International Energy Agency, "CO2 Emissions from Fuel Combustion - Highlights,"
IEA 2017, accesed on 10 August 2018,
https://www.iea.org/publications/freepublications/publication/CO2EmissionsfromFuel
CombustionHighlights2017.pdf.

[3] Government of Canada, "Regulations Limiting Carbon Dioxide Emissions from Natural
Gas-fired Generation of Electricity," Canada Gazette, 2018, accessed on 10 August
2018, http://gazette.gc.ca/rp-pr/p1/2018/2018-02-17/html/reg4-eng.html.

[4] Environment and Climate Change Canada, "Technical Backgrounder: Proposed

Federal Regulations for Electricity Sector," Government of Canada, 2018, accessed on
10 August 2018, https://www.canada.ca/en/environment-climate-
change/news/2018/02/technical_backgrounderproposedfederalregulationsforelectricity
se.html.

[5] M. Alvarez Cuenca, E. J. Anthony, Pressurized Fluidized Bed Combustion, 1st edition,
Springer Science and Business Media, Dordrecht 1995.

[6] M. Hirota, Japan Soc. Mech. Eng. Int. J. 2004, 47, 193.

[7] G. S. Grasa, M. Alonso, J. C. Abanades, Ind. Eng. Chem. Res. 2008, 47, 1630.

[8] C. Luo, Y. Zheng, J. Guo, B. Feng, Fuel 2014, 127, 124.

[9] B. Duhoux, P. Mehrani, D. Y. Lu, E. J. Anthony, R. Symonds, A. Macchi, Energy Tech.

2016, 4, 1158.

[10] R. T. Symonds, D. Y. Lu, R. W. Hughes, E. J. Anthony, A. Macchi, Chem. Eng. Sci.

2009, 64, 3536.

[11] R. T. Symonds, D. Y. Lu, A. Macchi, R. W. Hughes, E. J. Anthony, Chem. Eng. Sci.

2017, In Press.

[12] A. Perejon, J. Miranda-Pizarro, L. A. Perez-Maqueda, J. M. Valverde, Energy 2016,

 113, 160.

[13] S. K. Bhatia, D. D. Perlmutter, AIChE J. 1983, 29, 79.

[14] C. Ortiz, R. Chacartegui, J. M. Valverde, J. A. Becerra, L. A. Perez-Maqueda, Fuel

2015, 160, 328.

[15] R. T. Symonds, S. Champagne, F. Ridha, D. Lu., Powder Technol. 2016, 290, 124.

[16] S. Champagne, D. Y. Lu, A. Macchi, R. T. Symonds, E. J. Anthony, Ind. Eng. Chem.

Res. 2013, 52, 2241.

[17] F. Donat, N. H. Florin, E. J. Anthony, P. S. Fennell, Environ. Sci. Technol. 2012, 46,
1262.

[18] V. Manovic, E. J. Anthony, Energy & Fuels 2010, 24, 5790.

[19] J. Koike, S. Nakamura, H. Watanabe, T. Imaizumi, "Manufacturing and construction,

operation of Karita PFBC 360 MW unit," 17th International Conference on Fluidized
Bed Combustion, American Society of Mechanical Engineers, Jacksonville, 15 June
2003

[20] US DOE - NETL, "Tidd PFBC demonstration project: Project performance summary,"
DOE, 1999, accessed on 10 August 2018,
https://www.netl.doe.gov/File%20Library/Research/Coal/major%20demonstrations/cct
dp/Round1/TIDD/TIDD1.pdf

[21] M. B. Huleatt, "Handbook of Australian black coals : geology, resources, seam

properties, and product specifications," Australian Government, 1991, accessed on 10
August 2018, https://data.gov.au/dataset/572b6e69-1489-454b-b0c2-f6cdf5bac6dc

[22] US DOE - NETL, "Quality guidelines for energy system studies - Process modeling
design parameters", DOE/NETL, 2014, accessed on 10 August 2018
https://www.netl.doe.gov/energy-
analyses/temp/QGESSProcessModelingDesignParameters_042514.pdf.

[23] R. Turton, R. C. Bailie, W. B. Whiting, J. A. Shaeiwitz, D. Bhattacharyya, Analysis,

Synthesis, and Design of Chemical Processes, 4th edition, Prentice Hall, Boston, 2012.

[24] G. S. Grasa, J. C. Abanades, Ind. Eng. Chem. Res. 2006, 45, 8846.
[25] J. C. Abanades, Chem. Eng. J .2002, 90, 303.

[26] N. Y. Nsakala, G. N. Liljedahl, "Greenhouse gas emissions control by oxygen firing in

circulating fluidized bed boilers: Phase 1 - A preliminary systems evaluation," Alstom
Power Inc., 2003, accessed 10 August 2018, https://www.osti.gov/servlets/purl/825796.

[27] M. Broda, A. M. Kierzkowska, C. R. Müller, Environ. Sci. Technol. 2012, 46, 10849.

[28] Y. Du, Y. Yuan, G. T. Rochelle, Chem. Eng. Sci .2016, 155, 397.

[29] R. Chatti, A. K. Bansiwal, J. A. Thote, V. Kumar, P. Jadhav, S. K. Lokhande,

Microporous Mesoporous Mater. 2009, 121, 84.

[30] L. M. Romeo, Y. Lara, P. Lisbona, J. M. Escosa, Chem. Eng. J. 2009, 147, 252.