In Silico Solvent and Process Design for Carbon Capture-PhD Updrage Seminar

5/24/2018 In Silico Solvent and Process Design for Carbon Capture-PhD Updrage Seminar

1/33

In SilicoSolvent and Process Design forCarbon Capture

Laboratory for Multiscale Systems

Abdul Qadir

School of Chemical and Biomolecular Engineering

University of Sydney


2/33

Contents

Background and motivation Objective

System Design and Optimization Approach

Challenges

Polar Perturbed Chain- Statistical Association FluidTheory (PPC-SAFT)

Process and Solvent Optimization

Molecular Mapping of Hypothetical Solvent Results

Part 2: Solar Assisted Post-Combustion CarbonCapture

Future Work


3/33

Background

Australia is the worlds largest coal exporter

The majority of Australias electricity is produced from coal fired power plants

Generation capacity ~ 28 GW

Electricity production ~ 170 TWh/a

Average generation efficiency: ~35% for black coal

Average CO2 emissions: 0.9 tonne CO2/MWh

Average annual CO2-emissions: ~ 170 Mtonne CO2/a

Carbon tax legislation started in July 2012

Australian Government, DFAT, Composition of trade Australia, 2012.

34%

17% 16%

7%6%

15%

2% 3%

0

50

100

150

200

250

Energy -Electricity

Energy -Stationary

energy

excludingelectricity

Energy -Transport

Energy -Fugitive

emissions

Industrialprocesses

Agriculture Waste Land Use,Land-UseChange

andForestry

Annualemissions

(MtCO2-e)

National Greenhouse Gas Inventory, netemissions by sector, year to December 2012

AUSTRALIAN NATIONAL GREENHOUSE ACCOUNTS, Quarterly Update ofAustralias National Greenhouse Gas Inventory, December Quarter 2012

Australias principal exports


4/33

CO2capture technologies

Three main capture technologies:- Oxy-combustion

- Pre-combustion

- Post-combustion

http://www.vattenfall.com/en/ccs/


5/33

Pre-combustion carbon capture

CO2 Concentration:50 %vol

Compression, Heating and Pumping Energy

- Capture rate: 90%- Capture purity: 98%


6/33

Objectives

Capturee

nergy

penalty

Technology development

Capture penalty is thatenergy consumed by thecapture plant that otherwisewould have been used forpower generation.

Objective:

To assess techno-economic feasibility of carbon capture as a low-emissioncoal technology, via

Process and solvent optimization.


7/33

Superstructure level 2 (techno-economic)

Superstructure level 1 (Technical)

(e.g. Energy-Capture superstructure)

Plant Level

(e.g. Capture plant)

Unit Operation Level

(e.g. Absorber)

Sub-optimal designs and operations Optimal designs and operations

Optimization

Molecular Level

(eg. Solvent

Design )Power Plant

Capture Plant

Solar Thermal Plant

Electricity Market

Carbon Market

Weather Dynamics

Economics(CAPEX,OPEX)

ENERGY -CAPTURE

Superstructure

Approach: System Design and Optimization

Optimal design and operation

Flexibility and control

Safety

Economics

Integrated Process Design for Dynamic Conditions


8/33

Challenges

8

Optimization Challenges

Process parameters are continuous while material properties arediscrete.

Solvent Property Data Availability

Limited thermodynamic data for mixture of solvents.

Accurate prediction of thermodynamic properties of mixtures of

solvents required.


9/33

Polar Perturbed Chain - Statistical Association FluidTheory (PPC-SAFT)

9

Polar Perturbed Chain -Statistical Association Fluid Theory(PPC-SAFT) is a physically based EoS which can accuratelymodel pure component and mixture thermodynamic databased on molecular characteristics and attractive potentials.

Seven PPC-SAFT parameters (m, , /k, AB, AB/k,, xp)

PPC-SAFT perturbations of residual Helmholtz energy of material (Olthof ,

2009)


10/33

Two Routes to Optimization: Route 1

Use continuous molecular structural parameters to predict properties

of a hypothetical solvent through Continuous Molecular Targeting-Computer-Aided Molecular Design(CoMT-CAMD method), therebyconstructing a continuous optimization space.

10

Advantages Disadvantages

Requires a less complexoptimization algorithm

Solvent parameters are hypothetical and requiremapping to a real pure solvent or mixture.

Faster convergence tooptimal point

In order to find a true optimum, solvent mapping hasto be performed at many local optima as the globaloptima may not have a close match.

Solvent mapping techniques are not accurate,especially when predicting solvent mixtures.

Transport properties can not be predicted usingSAFT and thus those of the base solvent areassumed.


11/33

Two Routes to Optimization: Route 2

Use the discrete properties of real solvents and continuous processparameters to formulate a mixed integer non-linear programming(MINLP) optimization problem.

11

Advantages Disadvantages

Accurate solvent propertydata (including transportproperties)

Slower convergence

Does not require solvent

mapping stage(inaccuracies in mappingprevented)

A more complex algorithm

is used

Global optimum can beused


12/33

Pre-Combustion Carbon Capture Process

12

Simplified Pre-Combustion Carbon Capture process


13/33

Process and Solvent Optimization

13

Variable Description

Reference Solvent/ Initial

Conditions

m Number of segments in molecule 4.05

() Segment diameter 3.19

/k (K) Dispersive attraction energy 231.67

AB Association volume 1.96

AB/k (K) Association energy 1445

(D) Dipole moment 0

xp Dipole fraction 0

Temperature (K) Solvent regeneration temperature 330

HP Flash Pressure (atm) Pressure in high pressure flash vessel 15

LP Flash Pressure (atm) Pressure in low pressure flash vessel 14

()+ + ,+ , : Electricity generation penalty for extraction of steam @ 245 o C and 3bar from LP steam turbine in power plant.

s.t.

process equat ions

Variable bound const raints

CO2 capture > 90%

PHPF > PLPF

Solvent and Process Opt imizat ion Object ive funct ion

Table 1: Decisio n Variables .


14/33

Assumptions

14

Assumption Optimization Route1 (CoMT-CAMD)

OptimizationRoute 2 (MINLP)

Binary InteractionParameters=0

2-B associationmodel

Transport properties

assumed of basesolvent

x


15/33

Optimization Procedure

15

Optimization procedure Route 1 (left) & Route 2 (right); Optimization algorithm implemented in MATLAB and process flow sheet run in ASPENPlus. Software connectivity via VBA.

Genetic Algorithm optimization solver used.

CoMT-CAMD MINLP


16/33

Molecular Mapping of Hypothetical Solvent (Route 1)

16

+ .

=

Molecular target ing object ive funct ions

Fig. : Molecular mapping of pure solvent to hypothetical solvent.


17/33

Molecular Mapping of Hypothetical Solvent (Route 1)

17

Fig. : Molecular mapping of binary mixture of solvents to hypothetical solvent.


18/33

Results (CoMT-CAMD)

18

Solvent AAD Ranking Specific Energy Duty

(kJ/kgCO2)

Aniline 1 399

Triethylene Gylcol 2 706

Tetraethylene Glycol 3 380

Diethylene Glycol 4 7081-pentanol 5 425

Table 2: Solvent mappin g of pure solv ents.

Solvent TA Ranking Specific Energy Duty

(kJ/kgCO2)

Aniline 1 399

Pentaethylene Glycol 2 428

Tetraethylene Glycol 3 380

Diethylene Glycol 4 708

N-butanol 5 412


19/33

Results MINLP Optimization

19

Solvent Specific Energy Duty(kJ/kgCO2)

Solvent X1351.4

Solvent X2 350.7

Solvent 1 Solvent 2 Specific Energy

Duty (kJ/kgCO2)

Solvent Y1 Solvent Z1344.6

Solvent Y2 Solvent Z2347.2

Specific

Energy DutyReduction

Pure Solvent

Binary Solvent Mixture

Solvent Specific Energy Duty

(kJ/kgCO2)

NMP (Optimized process only)376.4


20/33

Part 2: Solar Assisted Post-CombustionCarbon Capture


21/33

Solar Assisted Post-combustion Carbon Capture(SPCC)

21

steamthermal energy for regeneration comes fromthe steam cycle by extracting high quality

steam from the turbines.

solar energy can be used to fully or partiallyprovide the solvent regeneration energy.

www.gastechnology.com.au


22/33

Net System Benefits

22

Formulates the SPCC operationalparameters and region-dependentvariables into expected costs andrevenues, to assess the expectedrevenue stream.

The net revenue from a power plant fittedwith an SPCCwould consist of thegenerated electricity sold at price pelec

(from which the following costs are subtracted: fuel, solarplant, CO2 pumping and storage cost, and the carbon costsincurred from the CO2that is actually released.)

Marwan Mokhtar, Muhammad Tauha Ali, Rajab Khalilpour, Ali Abbas, Nilay Shah, Ahmed AlHajaj, Peter Armstrong, Matteo Chiesa*, Sgouris Sgouridis Solar-Assisted Post CombustionCarbon Capture Feasibility StudyApplied Energy (2011)

(,B ,() ,A ) +(,B ,() ,A ) 2 2().

, + . ..+

Abdul Qadir, Marwan Mokhtar, Rajab Khalilpour, Dia Milani, Anthony Vassallo, MatteoChiesa, and Ali Abbas, 'Potential for Solar-Assisted Post-Combustion Carbon Capture in

Australia',Applied Energy,111 (2013), 175-85.


23/33

Heat Integration

23

Without With

Heat Integration

PCC:Post-combustionCarbon Capture

PCC

PowerPlant

SolarField

FinancialIncentives

SolarCollector

Technology

Location

PCC SolarField


24/33

Locations

24


25/33

Model Variables

25

Subsidy

Carbon Tax/Credits

Renewable EnergyCertificates

Flat Plate Collector(FPC)

Evacuated TubeCollector (ETC)

CompoundParabolic Collector

(CPC)

Linear FresnelCollector (LFC)

Parabolic TroughCollector (PTC)

Solar Collector Technologies Financial Incentives


26/33

Solar Collector Technologies: With and Without HeatIntegration

26

Without Heat Integration With Heat Integration

Comparison of solar collector technologies with and without heat integration for Sydney.

0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

Solar Load Fraction

Netannualbenefits[M$/yr]

Sydney

FPC

LFC

CPC

ETCPTC

0 0.2 0.4 0.6 0.8 1

-20

-15

-10

-5

0

Solar Load Fraction

Netannualben

efits[M$/yr]

Sydney

FPC

LFC

CPC

ETC

PTC

Abdul Qadir, Marwan Mokhtar, Rajab Khalilpour, Dia Milani, Anthony Vassallo, MatteoChiesa, and Ali Abbas, 'Potential for Solar-Assisted Post-Combustion Carbon Capture in

Australia',Applied Energy,111 (2013), 175-85.


27/33

Conservative Scenario (~$12/tonne-CO2)

27

Comparison of financialincentives for the four locationsunder the conservative carbonprice scenario.

0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

Solar Load Fraction

Ne

tannualbenefits[M$/yr]

Location: Sydney

0 0.2 0.4 0.6 0.8 1

-20

-15

-10

-5

0

5

10

Solar Loa d Fraction

Ne

tannualbenefits[M$/yr]

Location: Townsville

0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

Solar Load Fraction

Netannualbenefits

[M$/yr]

Location: Melbourne

Base Case

subsidy

subsidy+CCsubsidy+REC

Abdul Qadir, Marwan Mokhtar, Rajab Khalilpour, Dia Milani,Anthony Vassallo, Matteo Chiesa, and Ali Abbas, 'Potential forSolar-Assisted Post-Combustion Carbon Capture in Australia',

Applied Energy,111 (2013), 175-85.


28/33

Current Carbon Price ($23/tonne-CO2) Scenario

28

Comparison of financialincentives for the fourlocations under the current

carbon price scenario.

0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

Solar Load Fraction


Location: Sydney

0 0.2 0.4 0.6 0.8 1

-20

-15

-10

-5

0

5

10

Solar Load Fraction



0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

Solar Load Fraction

Netannualbenefits

[M$/yr]

Location: Melbourne

Base Case

subsidy



Applied Energy,111 (2013), 175-85.


29/33

Current price + 5% Annual Increase Scenario

29

0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

Solar Load Fraction

Neta

nnualbenefits[M$/yr]

Location: Sydney

0 0.2 0.4 0.6 0.8 1

-20

-15

-10

-5

0

5

10

Solar Load Fraction

Neta

nnualbenefits[M$/yr]


0 0.2 0.4 0.6 0.8 1-20

-15

-10

-5

0

5

10

Solar Load Fraction

Netannualbenefits[

M$/yr]

Location: Melbourne

Base Case

subsidy


Comparison of financialincentives for the fourlocations under the current

carbon price scenario.


Applied Energy,111 (2013), 175-85.


30/33

Conclusion

SPCC is viable for current carbon tax scenario for a solar fraction of 0.2 ifREC are supplied.

With an annual carbon price increase of 5%, SPCC is viable even withoutRECs at low SF.

SPCC is economically viable at a carbon price lower than that necessary

for carbon capture alone ($44 vs $58)

Eligibility for RECs would greatly boost SPCC viability and at the sametime increase solar technology deployment.

30


31/33

Future Work

Extension of physical solvent optimization tochemical solvent optimization

Novel solvent generation using groupcontribution methods

Solvent screening for high temperaturetolerant solvents for direct solar thermalregeneration

31

Absorber

Lean solvent cooler

Rich/Lean

heat exchanger

Exhaust gas

Hot Rich

Liquid

CO2

For compression

Flue gas

CO2

knock out

drum


32/33

Acknowledgements

Advisory Committee Dr Ali Abbas

Prof Tony Vassallo

Dr Matteo Chiesa

Research Group

Dr Rajab Khalilpour

Manish Sharma

Forough Parvareh

32


33/33

Thank you

Thank You!

Q&A

In Silico Solvent and Process Design for Carbon Capture-PhD Updrage Seminar

Documents

silico solvent

capture purity

optimization optimal

main capture technologies

pumping energy capture

absorbersuboptimal designs

operations optimal designs

background australia