CO Purification Unit (CPU) Models · CPU Performance Models The process design of CPU depends on the design constraints of the CO 2 product that goes into the pipeline. There are

IECM Technical Documentation:

CO2 Purification Unit (CPU) Models

January 2019

IECM Technical Documentation:

CO2 Purification Unit (CPU) Models

Prepared by:

Hari C. Mantripragada

Edward S. Rubin

The Integrated Environmental Control Model Team

Department of Engineering and Public Policy

Carnegie Mellon University

Pittsburgh, PA 15213

www.iecm-online.com

Compiled in January 2019

Integrated Environmental Control Model - Technical Documentation Table of Contents • iii

Table of Contents CO2 Purification Unit 1

Objective ............................................................................................................1 CPU Performance Models .................................................................................1

Flue Gas Compressor and Drying Model ..............................................3 Partial Condensation Model ...................................................................4 Distillation Model ..................................................................................8

CO2 Compressor Model .......................................................................10 CPU Efficiency Factor .....................................................................................10 Order of Calculations .......................................................................................11 CPU Cost Model ..............................................................................................12 User-Specified Configuration ..........................................................................13

IECM Screens ..................................................................................................14 Case Studies .....................................................................................................15

References ........................................................................................................17

Integrated Environmental Control Model - Technical Documentation List of Figures • iv

List of Figures

Figure 1. CO2 purity requirements for different applications (NETL QGESS, 2014) ......................... 2

Figure 2. Schematic of a typical CPU process (NETL, 2010) .............................................................. 3

Figure 3. Variation of CO2 purity and recovery with pressure and flash temperature ......................... 6

Figure 4. Variation of CO2 recovery with pressure and flash temperature ........................................... 7

Figure 5. CO2 recovery and purity for condensation and distillation ................................................... 9

Figure 6. Regression equations for capital cost of CPU ..................................................................... 13

Figure 7. "Set Parameters" config screen for the CPU unit ................................................................ 14

Figure 8."Set Parameters" purification screen for the CPU unit, 99.99 percent purity case .............. 14

Figure 9. "Set Parameters" purification screen for the CPU unit, high-purity case ............................ 15

Figure 10. "Set Parameters" purification screen for the CPU unit, low-purity case ........................... 15

Figure 11. Sensitivity analysis results ................................................................................................. 16

Figure 12. Sensitivity analysis for the CPU case studies .................................................................... 16

Integrated Environmental Control Model - Technical Documentation List of Tables • v

List of Tables

Table 1. Regression coefficients for the partial condensation model ................................................... 5

Table 2. Regression equations for the combined condensation and distillation model ........................ 9

Table 3. Case studies to calculate CPU efficiency factor ................................................................... 11

Table 4. Relevant flow rates and capital costs of CPU from the 2010 DOE Oxy-Combustion Report

...................................................................................................................................................... 12

Table 5. Case study input assumptions and results ............................................................................. 16

Integrated Environmental Control Model - Technical Documentation Acknowledgements • vi

Acknowledgements

This work was supported by the National Energy Technology Laboratory (NETL). The authors also

acknowledge Dr. Haibo Zhai and Karen Kietzke for their coding work in incorporating the

performance and cost models into the Integrated Environmental Control Model (IECM). Any

opinions, findings, and conclusions or recommendations expressed in this material are those of the

authors alone and do not reflect the views of any agency.

Integrated Environmental Control Model - Technical Documentation CO2 Purification Unit • 1

CO2 Purification Unit

Objective

This documentation reports the performance and cost models of a carbon dioxide (CO2) purification

unit (CPU) used for oxy-combustion and some post-combustion CO2 capture technologies at coal-

fired power plants.

CPU is needed to increase the purity of CO2 product streams to meet appropriate pipeline standards.

Some of these standards are prescribed in the National Energy Technology Laboratory’s (NETL)

Quality Guidelines for Energy Systems Studies (QGESS) document on CO2 purity requirements, as

shown in Figure 1 (NETL, 2013). Based on the specifications, CO2, water (H2O), oxygen (O2), argon

(Ar), and nitrogen (N2) are considered to be important components for design. Only these

components are considered in developing the CPU performance and cost models. Since sulfur oxides

(SOX) and nitrogen oxides (NOX) are removed almost completely by the CO2 capture units, in

conjunction with the sulfur dioxide (SO2) polishing unit, these components are not explicitly

considered in the models.

CPU Performance Models

The process design of CPU depends on the design constraints of the CO2 product that goes into the

pipeline. There are four major components of a CPU, as listed below (NETL, 2010; Besong et al,

2013):

1) Flue gas compression and drying: The CO2-rich flue gas that needs to be purified is first

compressed to a pressure of approximately 30 bar. Most of the water vapor in the flue gas

condenses because of this compression. Water vapor is further removed by an additional

drying unit. If no further purification is needed, this dry CO2-rich flue gas can be further

compressed for pipeline transport. Depending on the inlet flue gas, the CO2 product purity

from this step could be between 80 to 90 percent, with all the other impurities (mainly Ar,

N2, and O2) remaining in the CO2 product.

2) Partial condensation: Most pipelines require further improvement in CO2 purity. The

compressed and dry gas from the above-mentioned step is cooled using external refrigeration

to a temperature close to the critical point of CO2 (-59°C). The cooled gas is then sent

through a flash unit where the majority of the CO2 is separated from other gases. The CO2


product now reaches a purity of around 95 percent, with the other gases (Ar, N2, and O2)

contributing to the remaining 5 percent. The purity and recovery of CO2 can be adjusted

varying the operating conditions, such as the flash temperature and the number of flash

stages. A typical design includes a two-stage flash, which has been modeled in this study

(Pipitone and Boland).

3) Distillation: Though a high enough CO2 purity is achieved, the amount of O2 in the CO2

product cannot be adequately controlled with partial condensation. A further distillation step

is required to increase the CO2 purity to more than 99 percent and reduce the O2 content to

less than 100 parts per million (ppm).

4) CO2 final product compressor: The CO2 product from any of the above three steps is

compressed to pipeline pressure.

Figure 1. CO2 purity requirements for different applications (NETL QGESS, 2014)


Figure 2. Schematic of a typical CPU process (NETL, 2010)

These four steps have been modeled using Aspen Plus and the performance results are used to

develop reduced-order models (ROMs) for different equipment. The details of these models are

described below.

Flue Gas Compressor and Drying Model The first step in a CPU is the compression of flue gas (clean of impurities such as SOX and NOX) to

approximately 30 bar. In the process, most of the water vapor is condensed. The remaining water

vapor is removed in a separate drying unit.

A six-stage intercooled compressor was modeled using Aspen Plus (v9). Inlet water vapor

concentration and outlet pressure were varied to study their effect on the amount of water vapor

condensed in the compressor, the required compression energy and cooling duty for intercooling

between stages. ROMs are then developed for these variables. The isentropic efficiency of the

compressor is assumed to be 100 percent so that the Integrated Environmental Control Model

(IECM) user-defined efficiency can be used for calculation of the energy requirements. The

temperature of interstage cooling was fixed at 38°C, pressure was varied from 25 to 35 bar, and inlet

flue gas temperature was varied from 25 to 75°C.

The following regression equations were obtained:

EffH2O,condensed = 0.96362 + 0.000657*Pcompr1 + 0.08350*yH2O,fg – 0.000006*(Pcompr1)2 –

0.057983*(yH2O,fg)2 - 0.000482(Pcompr1)*( yH2O,fg) (R

2 = 98.48%) (1)

Q_cool_MJ/k = 5.182 + 0.0964*Pcompr1 – 0.0787*Tfg + 5.11*yH2O,fg + 0.003476*(Tfg)2 –


23.21*(yH2O,fg)2 (R2=95.27) (2)

EffQcool = 3.537 – 3.325*Effcompr – 0.02473*Tfg + 0.855*(Effcompr)2 + 0.000035*(Tfg)

2 +

0.02145*Effcompr*Tfg (R2 = 99.19%) (3)

Qcool,act (MJ/kmol inlet FG) = Qcool,ideal*EffQcool (4)

Wcompr,ideal (kWh/kmol inlet FG) = 1.93661 + 0.017046*Pcompr1 + 0.003314*Tfg – 2.59760*yH2O,fg

(R2=99.66) (5)

Wcompr,act = Wcompr,ideal/Effcompr (6)

Mole flow rate of condensed water is given as:

MH2O,condensed = EffH2O,condensed*MH2O,in (7)

All of the remaining water vapor in the dry gas is assumed to be dried in the dryer unit, which

utilizes activated alumina adsorbents. The adsorbents are regenerated using waste N2 streams. The

captured water vapor is assumed to go out with the waste N2 stream and is not accounted for in the

CPU model.

Mfg,dry = Mfg,in – MH2O,in (8)

This gas goes into the partial condensation unit.

Partial Condensation Model The compressed flue gas is sent to a partial condensation system. The dried flue gas is first cooled in

a heat exchanger to -27°C and then sent to a flash chamber where most of the CO2 condenses, along

with some other gases. The non-condensed gases, which consist of the remaining CO2 and other

gases, is sent to another heat exchanger where they are cooled to a much lower temperature (between

-59 and -40°C) and flashed again in a second flash chamber.

More CO2 is recovered from the gases. The condensed CO2-rich liquids from both the flash

chambers are mixed and either compressed to pipeline pressure or sent for further purification.

The partial condensation process was also modeled using Aspen Plus. To develop ROMs, the

following parameters were varied:

• Inlet flue gas composition

• Inlet pressure

• The second flash stage temperature

ROMs were developed for the purity and recovery of different gas components and for the cooling

load required for the process.

Inlet flue gas composition is expressed in terms of the ratio of molar flow rates (and hence mole

fractions) of different gas components to the CO2 mole flow. The IECM oxy-combustion module

was used to estimate the ranges of gas compositions by varying the coal type and excess air ratio.


The ranges for different components in the inlet flue gas are:

O2/CO2: 0.0077 – 0.1728

N2/CO2: 0.0998 – 0.1107

Ar/CO2: 0.0012 – 0.0603

Regression models were developed for the desired parameters. Table 1 shows the coefficients for

linear regression of mole fractions of CO2, Ar, and O2. Mole fraction of N2 is calculated as (1 – mole

fractions) of other components. The table also shows the coefficients for quadratic regression of

cooling duty and recovery of CO2 from the partial condensation unit.

yN2,out = 1 – yCO2,out – yN2,out – yO2,out.

Table 1. Regression coefficients for the partial condensation model

Coefficient yO2,out yAr,out yCO2,out recCO2 Qcool

(MJ/kmol CO2 prod)

Constant -0.02463 -0.00992 0.98136 -0.05332 54.537

Pcompr (bar) 0.000905 0.000356 0.0017 0.029831 -1.5628

Tflash (oC) -0.0003 -0.00011 0 -0.026195 0.71448

(Ar/CO2) -0.06356 0.190291 -0.17982 -1.9725 66.877

(N2/CO2) -0.07476 -0.0284 -0.09642 -2.196 72.484

(O2/CO2) 0.158934 -0.0231 -0.19083 -1.9456 67.805

(Pcompr)2 -7.40E-05 -0.000196 0.017315

(Tflash)2 0 -0.00017 0.004715

(Ar/CO2)2 0.10628 0 4.71

(N2/CO2)2 -0.22376 0.0989 4.66

(O2/CO2)2 0.10935 -0.0572 6.618

(Pcompr)*(Tflash) 0 0.000342 -0.010259

(Pcompr)*(Ar/CO2) 0.001954 0.01509 -0.8547

(Pcompr)*(N2/CO2) 0.00594 0.014913 -0.8725

(Pcompr)*(O2/CO2) 0.001843 0.01498 -0.871

(Tflash)*(Ar/CO2) 0 -0.022648 0.6409

(Tflash)*(N2/CO2) 0 -0.025147 0.7117

(Tflash)*(O2/CO2) 0 -0.022445 0.63959

(Ar/CO2)*(N2/CO2) -0.06733 0 11.36

(Ar/CO2)*(O2/CO2) 0.20252 -0.1101 12.92

(N2/CO2)*(O2/CO2) -0.0512 0 12.93

R2 96.02 94.72 99.74 99.5 98.7

From the regression equation, it was found that purity varies more with pressure than with

temperature. At each pressure, the range of variation of purity with temperature is 0.008. An average

temperature of -50°C is fixed for the purity equation. By doing this, purity is now a function of flue

gas composition and pressure only. The adjusted coefficients are given in the table.


The user fixes purity and recovery. From these two values, pressure and temperature are calculated.

Figure 3. Variation of CO2 purity and recovery with pressure and flash temperature

Range of CO2 Purity: For a given temperature, purity increases with lower pressures. Therefore, the

minimum purity occurs at 35 bar and maximum purity occurs at 25 bar. These values are used to

calculate the range of CO2 purity.

yCO2_min = yCO2 at 35 bar (9)

yCO2_max = yCO2 at 25 bar. (10)

For a specified purity, the operating pressure is calculated by solving the quadratic equation of CO2

purity.

Quadratic term of the equation is calculated as follows:

a_p=a_yCO2(7);

Linear term of the equation is calculated as follows:

b_p=a_yCO2(2) + a_yCO2(13)*(Ar/CO2) + a_yCO2(14)*(N2/CO2) + a_yCO2(15)*(O2/CO2);

Constant term of the equation is calculated as follows:

c_p=a_yCO2(1) + a_yCO2(3)*0 + a_yCO2(4)*(Ar/CO2) + a_yCO2(5)*(N2/CO2) +

a_yCO2(6)*(O2/CO2) + a_yCO2(9)*(Ar/CO2)^2 + a_yCO2(10)*(N2/CO2)^2 +

a_yCO2(11)*(O2/CO2)^2 + (Ar/CO2)*(a_yCO2(19)*(N2/CO2) + a_yCO2(20)*(O2/CO2)) +

a_yCO2(21)*(N2/CO2)*(O2/CO2) – yCO2

Finally, pressure is calculated as follows:

p = (- b_p – sqrt(b_p^2-4*a_p*c_p))/(2*a_p) (11)


Figure 4. Variation of CO2 recovery with pressure and flash temperature

Range of CO2 Recovery: At a given pressure, recovery increases with lower temperature. Therefore,

the minimum recovery occurs at -40°C and maximum recovery occurs at -55°C. These values are

used to calculate the range of CO2 recovery, for the pressure calculated above.

recCO2_min = recCO2 at -40°C. (12)

recCO2_max = recCO2 at -55°C. (13)

For a specified pressure (calculated from CO2 purity) and recovery, the operating temperature is

calculated by solving the quadratic equation of CO2 recovery.

Quadratic term of the equation is calculated as follows:

a_T_flash = a_recCO2(8);

Linear term of the equation is calculated as follows:

b_T_flash = a_recCO2(3) + a_recCO2(12)*p + a_recCO2(16)*(Ar/CO2) + a_recCO2(17)*(N2/CO2)

+ a_recCO2(18)*(O2/CO2);

Constant term of the equation is calculated as follows:

c_T_flash = a_recCO2(1) + a_recCO2(2)*p + a_recCO2(4)*(Ar/CO2) +

a_recCO2(5)*(N2/CO2) + a_recCO2(6)*(O2/CO2) + a_recCO2(7)*p^2 +

a_recCO2(9)*(Ar/CO2)^2 + a_recCO2(10)*(N2/CO2)^2 + a_recCO2(11)*(O2/CO2)^2 +

p*(a_recCO2(13)*(Ar/CO2) + a_recCO2(14)*(N2/CO2) + a_recCO2(15)*(O2/CO2)) +

(Ar/CO2)*(a_recCO2(19)*(N2/CO2) + a_recCO2(20)*(O2/CO2)) +

a_recCO2(21)*(N2/CO2)*(O2/CO2) – recCO2.

Finally, temperature is calculated as follows:

T_flash = (-b_T_flash – sqrt(b_T_flash^2-4*a_T_flash*c_T_flash))/(2*a_T_flash)

(14)


Flow rates of components in the CO2 product are given as:

MCO2,product = recCO2*MCO2,in

MN2,product = yN2,out*MCO2,product/yCO2,out

MO2,product = yO2,out*MCO2,product/yCO2,out

MAr,product = yAr,out*MCO2,product/yCO2,out

Distillation Model Distillation is used to further purify CO2. The main design parameter for the distillation column is

the amount of O2 present in the CO2 product. Depending on the application, O2 purity should be <

100 ppm or < 10 ppm.

recCO2 = 0.88616 – 0.21720*((Ar+N2)/CO2) – 0.27630*(O2/CO2) + 3.8x10-5*(O2 ppm) –

0.0001812*(Tdistill) (R2=98.97%)

If O2 ppm and Tdistill are specified, recCO2 can be calculated from the equation above. On the other

hand, if O2 ppm and recCO2 are specified, the above equation can be used to calculate Tdistill.

To simplify calculations for a combined partial condensation and distillation process, Tdistill is fixed

at its minimum value of -59°C and O2 concentration is fixed at 10 ppm.

Qcool,distill (MJ/kmol gas in) = 135.09 – 287.89*yCO2,in – 0.078406*Tdistill + 153.38*( yCO2,in)2

(R2=98.97%)

The CO2 product from the distillation column is almost 100 percent CO2 and a very small amount of

O2. The presence of N2 and Ar is negligible, compared to the concentration of O2. Hence, it is

assumed that only CO2 and O2 are present in the CO2 product and the other gases go out through the

vent stream.

Molar flow rates of gases in CO2 product and vent streams are calculated as follows:

MCO2,product = recCO2*MCO2,in

MO2,product = yO2/(1 – yO2)*MCO2,product

MCO2,vent = (1-recCO2)*MCO2,in

MO2,vent = MO2,in – MO2,product

MN2,vent = MN2,vent

MAr,vent = MAr,vent

For simplifying calculations, a combined partial condensation and distillation model was developed.

Inputs to this are the flue gas composition into CPU.

Table 2 shows the coefficients for parameters in a combined partial condensation and distillation

model.


Table 2. Regression equations for the combined condensation and distillation model

Coefficient recCO2

Qcool (distillation

only)

(MJ/kmol CO2 prod)

Coefficient T_flash (°C)

Constant -0.038805 3.11854 Constant -346.74

Pcompr (bar) 0.028721 -0.04888 Pcompr (bar) 6.113

Tflash (oC) -0.025767 0.013564 (Ar/CO2) -280.4

(Ar/CO2) -1.951910 -2.44382 (N2/CO2) -325.3

(N2/CO2) -2.160660 0.737105 (O2/CO2) -281.94

(O2/CO2) -1.933630 -2.7221 recCO2 766.8

(Pcompr)2 -0.000195 0.000554 (Pcompr)

2 -0.00918

(Tflash)2 -0.000168 4.30E-05 (Ar/CO2)

2 166.7

(Ar/CO2)2 0.005637 1.37139 (N2/CO2)

2 229.9

(N2/CO2)2 0.091358 0.12884 (O2/CO2)

2 161.6

(O2/CO2)2 -0.045878 1.73731 (recCO2)

2 -498.36

(Pcompr)*(Tflash) 0.000339 -0.00031 (Pcompr)*(Ar/CO2) -1.631

(Pcompr)*(Ar/CO2) 0.014839 0.055146 (Pcompr)*(N2/CO2) -2.179

(Pcompr)*(N2/CO2) 0.014942 -0.0169 (Pcompr)*(O2/CO2) -1.7526

(Pcompr)*(O2/CO2) 0.014618 0.062067 (Pcompr)*(recCO2) -4.888

(Tflash)*(Ar/CO2) -0.022148 -0.01531 (Ar/CO2)*(N2/CO2) 375.1

(Tflash)*(N2/CO2) -0.024796 0.004694 (Ar/CO2)*(O2/CO2) 317.63

(Tflash)*(O2/CO2) -0.021849 -0.01724 (Ar/CO2)*(recCO2) 177.3

(Ar/CO2)*(N2/CO2) 0.005347 -0.84045 (N2/CO2)*(O2/CO2) 383.01

(Ar/CO2)*(O2/CO2) -0.091822 3.08658 (N2/CO2)*(recCO2) 216

(N2/CO2)*(O2/CO2) 0.013744 -0.94617 (O2/CO2)*(recCO2)* 180.5

R2 100 98.7 R2 94.31

T_flash prediction was close to temperatures higher than -55°C. So the lower end of T_flash is

changed from -59°C to -55°C. The range of T_flash now is -55°C to -40°C.

Figure 5. CO2 recovery and purity for condensation and distillation


Recovery is higher at lower temperatures and higher pressures. For a given flue gas,

recCO2_min = recovery at 25 bar and -40°C (15)

recCO2_max = recovery at 35 bar and -55°C (16)

Fix p=30 bar. If T_flash is <-55°C, then p=35 bar. If T_flash>-40°C, then p=25 bar. (17)

CO2 Compressor Model The CO2 compressor was also modeled in Aspen Plus as a function of varying inlet and outlet

pressures. Inlet pressure was varied between 25 and 35 bar, and the outlet pressure was varied

between 120 and 160 bar. The following regression equations were developed from the results of the

Aspen models.

Work required for compression (kWh/kmol CO2 product total):

WCO2compr,ideal (kWh/kmol CO2 product total) = 1.12905 – 0.023379*Pin(bar) + 0.001975*Pout(bar)

(R2=99.57%)

WCO2compr,act (kWh/kmol CO2 product total) = WCO2compr,ideal/EffCompr (18)

WCO2compr,act (kWh/kmol CO2) = WCO2compr,act (kWh/kmol CO2 product total) / Purity

Cooling duty required for compressor inter-cooling is given by the following equations:

Qcool,CO2compr (MJ/kmol CO2 prod total) = 15.4410 – 0.14584*Pin(bar) + 0.026112*Pout(bar) –

3.3577*Eff_compr (R2=99.59) (19)

To convert it to per kmol of CO2, this has to be divided by CO2 purity. This heat is supplied by

cooling water.

CPU Efficiency Factor

The CPU refrigeration load requirement (partial condensation and distillation) calculated using the

models above were validated with the energy requirement values presented in the DOE 2010 oxy-

fuel report. The DOE reports present the total CPU energy requirement, which includes refrigeration

and compression. For these case studies, compression work is calculated from the IECM models

developed above and subtracted from the total work given in the DOE report. KeyLogic clarified

that the refrigeration system used is an auto-thermal system, which means that the refrigeration load

is met by internal heat integration without any need for external electrical input. A CPU adjustment

factor was calculated to account for modeling differences between the Aspen models and the models

used in the DOE report. The coefficients are shown in Table 3.


CPUfactor = WDoE/WIECM (20)

CPUfactor = 1.327 for compression only and 1.054 for partial condensation (21)

CPUfactor = 1.102 for partial condensation + distillation (22)

The overall CPU energy requirement is calculated as:

WCPU (kWh/kmol CO2)= (WCompr,fg + WCO2Compr)*CPUfactor (23)

Units for all the parameters above should be converted to kWh/kmol CO2.

Table 3. Case studies to calculate CPU efficiency factor

DOE case S12E S12F L12F S13F L13F

CPU configuration Partial

condensation Partial condensation + Distillation

Inlet CO2 mole fraction 0.6994 0.6994 0.708 0.6997 0.6871

Inlet H2O mole fraction 0.1726 0.1726 0.1627 0.1726 0.1726

Ar/CO2 0.0419 0.0419 0.0415 0.0419 0.0426

N2/CO2 0.1090 0.1090 0.1086 0.1089 0.1093

O2/CO2 0.0320 0.0320 0.0323 0.0316 0.0521

Temperature of flue gas (°C) 57 57 57 57 57

Purity 0.9771 0.9998 0.9998 0.9998 0.9998

Recovery 0.9102 0.909 0.909 0.9116 0.909

CO2 product total (kmol/hr) 12,285 12,049 12,707 11,543 11,601

CPU energy total (kW) 62,100 64,740 67,140 62,090 65,780

CPU energy (kWh/kmol CO2 prod) 5.17 5.37 5.29 5.38 5.67

Calculated Values

Pressure (bar) 25.0 30.0 30.0 30.0 30.0

T_flash (°C) -47.9 -46.7 -46.6 -47.3 -48.8

Refrigeration load (kWh/kmol CO2 prod) 5.08 5.29 5.29 5.27 5.31

Compression load (kWh/kmol CO2 prod) 4.91 4.91 4.91 4.90 4.98

CPU efficiency factor 1.054 1.095 1.077 1.099 1.138

Order of Calculations

The following order should be used for the overall system calculation:

1. Choose a configuration:

a. Capture with impurities (only flue gas compression and drying).

b. High CO2 purity (95 to 97 percent) (partial condensation).

c. Ultra-high CO2 purity (99.99 percent) (partial condensation + distillation).

2. Specify CO2 purity and recovery.


a. Config (a): Purity = CO2 mole fraction after drying.

b. Config (b): User specifies purity and recovery. Ranges are calculated using Equations

9, 10, 12. and 13.

c. Config (c): Purity is fixed at 99.99 percent (O2 purity = 10 ppm). Ranges for recovery

are calculated using Equations 15 and 16.

3. Determine system operating pressure.

a. Config (a): Assume 30 bar.

b. Config (b): Equation 11, using coefficients from Table 1.

c. Config (c): Equation 17, using coefficients from Table 2.

4. Determine flash temperature.

a. Config (a): No flash.

b. Config (b): Equation 14, using coefficients from Table 1.

c. Config (c): Equation 17, using coefficients from Table 2.

5. Determine compression energy using Equation 6 and Equation 18.

6. Determine cooling duty to be supplied by cooling water using Equation 4 and Equation 19.

7. Determine refrigeration duty.

a. Config (a): No refrigeration.

b. Config (b): Qcool using coefficients in Table 1.

c. Config (c): Qcool using coefficients from Table 1 + Table 2.

8. Calculate CPU efficiency using Equation 21 or Equation 22.

9. Calculate total CPU refrigeration energy required using Equation 23.

CPU Cost Model

CPU capital cost data was obtained from the 2010 DOE report on oxy-combustion power plants.

Table 4 shows the data, along with relevant flow rates.

Table 4. Relevant flow rates and capital costs of CPU from the 2010 DOE Oxy-Combustion

Report

The costs for individual components are plotted and exponential regression equations were fit as a

function of flow rates. The capital cost of compression and drying unit fits well with CO2 product

flow rate, while that of the condensing heat exchanger fits well with the total CPU inlet flow rate, as

shown in Figure 6.

The final cost functions are given as:

DOE Oxy-Fuel Report Case S12D S12E S12F L12F S13F L13F

Total CPU inlet flow (kmol/hr) 20,281 18,885 18,977 19,772 18,712 18,598

Total CO2 prod flow (kmol/hr) 15,952 12,285 12,049 12,707 11,543 11,601

CO2 Removal Cost (x $1,000, 2007)

CO2 condensing HX $ 3,424 $ 4,292 $ 4,256 $ 3,759 $ 4,169 $ 4,376

CO2 compression and drying $ 98,224 $ 77,523 $ 77,806 $ 80,579 $ 75,697 $ 78,375

Total $ 101,648 $ 81,815 $ 82,062 $ 84,338 $ 79,866 $ 82,751


CCondensingHX ($M, 2007) = 2x1012 (CPU inlet flow, kmol/hr)– 2.714

CCompr and Drying ($M, 2007) = 0.05017 (Total CO2 product, kmol/hr)0.7824

Note: CPU inlet flow contains water vapor. Total CO2 product includes CO2 as well as other

impurities.

CCPU,total = CCondensingHX + CCompr and Drying

Figure 6. Regression equations for capital cost of CPU

Other than the internal electricity consumption, there are no other variable operations and

maintenance (VOM) costs for the CPU unit. Fixed operations and maintenance (FOM) costs are the

same as the standard IECM defaults.

User-Specified Configuration

In this configuration, users can specify their own purity, recovery, energy penalty, and cost. Mass

balance equations are given below.

Mgas,in (kmol/hr) = flue gas into CPU after direct contact cooler, sulfur polisher, and drying. (after

removal of H2O and all other components, other than Ar, N2, CO2, and O2).

MCO2,out (kmol/hr) = MCO2,in * Recovery

Mgas,out = MCO2,out/purity

MAr,out = (1/3)*(1 – purity)*Mgas,out

MN2,out = (1/3)*(1 – purity)*Mgas,out

MO2,out = (1/3)*(1 – purity)*Mgas,out

User can specify energy penalty in kWh/tonne CO2 (default value is 120 kWh/tonne) and


CPU capital cost in $M/tonne/hr CO2 (default value is 0.175).

Direct cost of CPU = (CPU unit cost, $M/tonne/hr CO2)*(CO2 product, tonne/hr).

IECM Screens

Figure 7 through Figure 10 show the input (Set Parameters) and output (Get Results) screens of the

CPU model in IECM. The plant configuration is a pulverized coal plant with a membrane-based

post-combustion CO2 capture unit.

Figure 7. "Set Parameters" config screen for the CPU unit

Figure 8."Set Parameters" purification screen for the CPU unit, 99.99 percent purity case


Figure 9. "Set Parameters" purification screen for the CPU unit, high-purity case

Figure 10. "Set Parameters" purification screen for the CPU unit, low-purity case

Case Studies

Case studies were conducted using the performance and cost models developed before. NETL case

S12E was used as the base case for inlet flue gas flow rate and composition. The results are


presented in Table 5. Sensitivity analysis was conducted by varying the purity and recovery. Results

are shown in the following figures.

Figure 11. Sensitivity analysis results

Figure 12. Sensitivity analysis for the CPU case studies

Table 5. Case study input assumptions and results

Configuration Compression Only Partial Condensation Distillation

Inlet CO2 mole fraction 0.6994 0.6994 0.6994

Inlet H2O mole fraction 0.1726 0.1726 0.1726

Inlet flue gas temperature (°C) 57 57 57

Ar/CO2 0.0419 0.0419 0.0419

N2/CO2 0.1090 0.1090 0.1090

O2/CO2 0.0320 0.0320 0.0320

CPU inlet flow (kmol/hr) 18,885 18,885 18,885

Dried gas flow (kmol/hr) 15,625 15,625 15,625

Purity (%) 84.5 95.11 99.99

Recovery (%) 100 92.07 90.32

CO2 product total (kmol/hr) 15,625 12,787 11,931

Minimum purity (%) 95.11 99.99


Configuration Compression Only Partial Condensation Distillation

Maximum purity (%) 97.06 99.99

Pressure (bar) 35.0 30.0

Minimum recovery (%) 92.07 84.32

Maximum recovery (%) 95.80 93.91

Flash temperature (°C) -40.0 -45.3

Refrigeration duty (kWh/kmol CO2 prod) 4.69 5.69

W_compr_total (kWh/kmol CO2 prod) 4.65 4.87 4.88

W_compr_total (kWh/tonne CO2) 105.5 110.5 110.6

eff_CPU 0.000 0.053 0.095

W_CPU_total (kWh/kmol CO2) 4.65 5.12 5.42

Cost_CondensHX ($M, 2007) 4.96 4.96 4.96

Cost_Compr and Drying ($M, 2007) 95.87 81.96 77.63

Cost_CPU_total ($M, 2007) 100.84 86.92 82.59

Cost_CPU_total ($M/tonne/hr CO2, 2007) 0.1735 0.1624 0.1573

References

Besong, M.T., Maroto-Valer, M.M., Finn, A.J., (2013). Study of design parameters affecting the

performance of CO2 purification units in oxy-fuel combustion. International Journal of

Greenhouse Gas Control. 12, 441-449.

National Energy Technology Laboratory (2010). Cost and performance for low-rank pulverized coal

oxy-combustion plants, Final Report, DOE/NETL-401/093010.

National Energy Technology Laboratory (2013). Quality guidelines for energy system studies – CO2

impurity design parameters, DOE/NETL-341/011212.

CO Purification Unit (CPU) Models · CPU Performance Models The process design of CPU depends on the design constraints of the CO 2 product that goes into the pipeline. There are

Documents