Emad Abbasi, Javad Abbasian and Hamid Arastoopour

Motivation and Background

Objective, Scope and Timeline

Completed Work and Results

Future Work

Questions?

petroleum 3%

Coal 51%

Natural gas 16%

Non fossil 30%

Electricity sector energy consumption 2008

petroleum 6% Coal

53%

Natural gas 25%

Non fossil 16%

Electricity sector energy consumption 2030

petroleum 38%

Coal 23%

Natural gas 24%

Non fossil 15%

US Primary Energy Consumtion 2008

petroleum 4%

Coal 81%

Natural gas 15%

Resulting CO2 emission by Electricity sector

WGS Reaction

CO2 Removal

Fuel Gas

CO2

Conventional

WGS Reaction &

CO2 Removal

Fuel Gas

CO2

T: 350°-500°C P: 10-70 bar

Integrated

Concentrated Hydrogen Stream

Hydrogen Polishing

Fuel Cells, Transportation

Impurities

Chemical Syn./ Liquid Fuels

Concentrated Hydrogen Stream

WGS Reaction &

CO2 Removal

CO + H2O CO2 + H2

XO + CO2 XCO3

Clean Coal Gas

Concentrated Hydrogen Stream

Sorbent Regeneration

XCO3 XO + CO2

CO2

XO make up

0.01

0.1

1

10

100

250 300 350 400 450 500Temp, C

P CO

2, at

m

AbsorptionMgO+CO2 MgCO3

DecompositionMgCO3 MgO+CO2

°

Advanced Power Plant

The overall objective of this project is to develop a CFD model and to perform Computational Fluid Dynamic (CFD) simulations using Population Balance Equations (PBE) to describe the heterogeneous gas-solid absorption/regeneration and water-gas-shift (WGS) reactions in the context of multiphase CFD for a regenerative magnesium oxide-based (MgO-based) process for simultaneous removal of CO2 and enhancement of H2 production in coal gasification processes.

The Project consists of the following four (4) tasks: Task1. Development of a CFD/PBE model accounting for the particle

(sorbent) porosity distribution and of a numerical technique to solve the CFD/PBE model.

Task2. Determination of the key parameters of the absorption and regeneration and WGS reactions.

Task3. CFD simulations of the regenerative carbon dioxide removal

process. Task4. Development of preliminary base case design for scale up

Task 1.

Task 2.

Task 3.

Task 4.

Month 3 6 9 12 15 18 21 24 27 30 33 36

Milestones:Task completionExperimental work completedReaction model finalized CFD simulation of single reaction/reactor CompletedCFD simulation of integrated process CompletedDevelopment of the base-case design completed

Carbonation:

CO2 + H2O+ K2CO3 → 2KHCO3 + Heat

Regeneration:

2KHCO3 → CO2 + H2O+ K2CO3 - Heat

-------------------------------------------------------------------------------------------------------------------------------------------------------------- Yi et al., International journal of greenhouse gas control, 2007.

2D, Eulerian- Eulerian Approach in combination with the kinetic theory of granular flow

Assumptions: Uniform and constant particle size and density - Conservation of Mass - gas phase: - solid phase - Conservation of Momentum - gas phase: - solid phase

gggggg mvt

•

=∇+∂∂ ).()( ρερε

ssssss mvt

•

=∇+∂∂ ).()( ρερε

)(.).()( sggsggggggggggg vvgPvvvt

−−+∇+∇−=∇+∂∂ βρετερερε

)(.).()( sggsssssssssssss vvgPPvvvt

−++∇+∇−∇−=∇+∂∂ βρετερερε

Task 1

Numerical Modeling: Conservation Equations

- Conservation of Species - gas phase: - solid phase - Conservation of solid phase fluctuating Energy - solid phase

- Reaction Kinetic: Deactivation Kinetic Model (Park et al, 2007)

jisssiss Ryvyt

=∇+∂∂ ).()( ρερε

Generation of energy due to solid

stress tensor

Diffusion dissipation

ssssssssss vIpvt

γθκτθρεθρε −∇∇+∇+−∇=∇+∂∂ ).(:)(]).()([

23

jigggigg Ryvyt

=∇+∂∂ ).()( ρερε

)]exp()exp(1

)))]exp(1(.exp(1[exp[ tk

tktkk

a dd

ds −−−

−−−=

τ

aCkdtda

COd 2=−

Task 1

Numerical Modeling: Conservation Equations

Gas-solid inter-phase exchange coefficient: EMMS model (Wang et al. 2004)

Accounts for cluster formation by multiplying the “Wen & Yu” drag

correlation with a heterogeneity factor

Heterogeneity Factor

ω < 1 =sgβ

)()1(

43

0 gDsggp

gg Cuud

εωρεε

−−

p

sggg

pg

gg

d

uu

d

−−+

− ρε

ε

µε )1(75.1

)1(150 2

2

74.0>gε

74.0<gε

=)( gεω

0044.0)7463.0(40214.05760.0 2 +−

+−gε

0040.0)7789.0(40038.00101.0 2 +−

+−gε

gε8295.328295.31 +−

82.074.0 ≤< gε

97.082.0 ≤< gε

97.0>gε

Task 1

Numerical Modeling: Drag Correlation

t: 1s 4s 8s 11s 14s 17s 20s 23s 26s 29s 33s

Time averaged Pressure drop

(mm H2O) KIER

Experiments

Time averaged Pressure

drop (mm H2O) Simulation

DP1 100 107 DP2 200-500 335 DP3 250 270 DP4 70 73

Pressure Drop

10

20

30

40

50

60

70

1 1.5 2 2.5 3 3.5

CO

2R

emov

al %

Inlet Gas Velocity (m/s)

Experiment

Simulation-Diactivation model

Simulation-Homogenous Model

Baseline operation condition

sensitivity to the Inlet Gas Velocity

Contours of CO2 mass fraction

Task 1

Results

-------------------------------------------------------------------------------------------------------------------------------------------------------------- Yi et al., International journal of greenhouse gas control, 2007.

What is the Population Balance Equation?

The population balance equation is a balance equation based on the number density function f (ξ; x, t)

Accounts for the particles accumulating, leaving, entering or being generated or destroyed in a single control volume

),;()],;([]),;(),;([)],;(),([),;( thtftx

tftDx

tftuxt

tf j

jipt

ip

i

xξxξxξxξxξxxξ=

∂

∂

∂∂

+∂

∂∂∂

+∂∂

+∂

∂ ξξ

Accumulation term + Convection term + diffusive term + Growth term = Source term

Finite size domain Complete set of trial functions Method Of Moments: FCMOM Finite size domain: [-1, 1] instead of [0,∞]

Solution in terms of both Moments and size distribution f(ξ,x,t) will be approximated by expansion based on a complete set of trial

functions

∑

∑

=−

−

∞

=

−−−

+=

Φ=

n

vnv

vnnn

nn

n

vvnnvvnc

xtCtxf

02

0

}.])!].[()![(

1.{])!2[(

)!2(.)1(.21.

212

when )().,(),,(

µ

ξξ

ξξµ df ii .).(

1

1∫−

′= )(.2

12)( ξξφ nn Pn +=

2/)]()([}2/)]()([{

maxmin

maxmin

tttt

ξξξξξ

ξ++−

=

)().()..( 321 IGMBMBMBMBMBDvt DiffDiffDiffConviptpi

i +++++−=∇′∇−∇+∂∂

µµµ

MB : Terms due to coordinate transformation (Moving Boundary)

IG: Contribution due to the Integration of Growth Term

maxmax

minmin and S

dtd

Sdt

d==

ξξ

Boundary conditions:

)()..( IGMBMBvt Convpi

i ++−=∇+∂∂

µµ

For the application of interest:

Uniform and constant particle size distribution.

Density of the particles is changing during the process due to the reaction between the solid and the gas phase.

Density distribution function is defined in the range of [ξmin, ξmax] and then using a coordinate transform is changed to [-1, +1].

Incompressible particle phase .

Constant maximum sorbent density, corresponding to the completely reacted sorbent.

Variable minimum sorbent density, corresponding to the fresh sorbent. The rate of change is related to the rate of reaction.

no breakage or agglomeration in density domain.

Implementation in Ansys Fluent via UDS is

is

iss

ispss

isss SDv

t φφεφρεφρε

=∇−∇+∂

∂).(

s

iis ε

µφ =

)()..( IGMBMBvt Convpi

i ++−=∇+∂∂

µµ

Test case1: Linear Growth, No convection

Kt i

i .µµ

=∂∂

Case study for k=0.05 and t=10 sec

1− 0.5− 0 0.5 15−

0

5

10

15

Particle Property (Dimensionless)

Num

ber d

ensit

y fu

nctio

n (D

imen

sionl

ess) 11.333

0.011−

I ξ( )

A ξ( )

f10 ξ( )

11− ξ1− 0.5− 0 0.5 1

5−

0

5

10

15

Particle Property (Dimensionless)

Num

ber d

ensit

y fu

nctio

n (D

imen

sionl

ess) 11.33

3.345−

I ξ( )

A ξ( )

f6 ξ( )

11− ξ

6 moments 10 moments

CFD

Two Fluid Model

Phase velocity, Volume Fraction

Mean Solid Density

UDS Source terms Moments of Density Distribution

Population Balance Model

Reaction

Assumption: Moments are convected with mixture velocity

Kvt p =∇+∂

∂ . minmin ξ

ξ

).()(

}.).1(].)1({[

).()(

}..].)1({[

).()(

1}.).1(].)1({[

).()(

1}..].)1({[ ][

min

minmax

,1

11

min

minmax

,111

min

minmax1

11

min

minmax111,

j

jpi

i

j

jpi

i

ii

ii

ijpj

i

xv

iff

xv

iff

dtdiff

dtdiffv

xt

∂∂

−−

+−′−−′

−∂∂

−−′−−′

−−−

+−′−−′

−−

−′−−′−=∂∂

+∂∂

−+

+

−−+

−+

+

−−+

ξξξ

µ

ξξξ

µ

ξξξ

µ

ξξξ

µµµ

2

)())(( maxminminmax0

1 ξξξξµµ

ρ++−

=s

vg= vs= 1 m/s

εs = 0.2

f2 f

f1

1m

1st Moment 2nd Moment

1− 0.5− 0 0.5 1

0

5

10

15

Particle Property (Dimensionless)

Num

ber d

ensit

y fu

nctio

n (D

imen

sionl

ess)

I ξ( )

f ξ( )

f1 ξ( )

f2 ξ( )

ξ

t = 10s

t = 10s

Minimum Density

Solid Volume Fraction

Solid Density

* A. Hassanzadeh, 2007

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30Time, min

Con

vers

ion,

%

20 µm Scale:

k2

k1

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

k, (c

m/m

in)

Radius of Particle (μm)

k2

k1

Potassium promoted MgO-based sorbent

1- There are two distinct reactive zones inside the particles 2- Process is controlled by both surface reaction and product

layer diffusion 3- There is an Expanding product layer 4- Dg is Variable due to the pore closing and is a function of

conversion 5- Intrinsic reaction rate is Arrhenius type

)exp(0 RTEkk ss −=

)(0βαXDD gg −=

3 )1( ZXXrr pp +−′=

rc : Radius of the low reactive zone (k2) rp : Initial radius of the particles rp’: Radius of the expanded particle Low Reactive Zone (k2)

Highly Reactive Zone (k1)

Product Layer

Gas Film

rp, k1

rc k2

rp’

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0 200 400 600 800 1000 1200 1400 1600

MgO

Con

vers

ion

Time (s)

Exp 425°C

Variable Dg

SCM

action

s

Layeroduct

gp

iip

gp

i

oMgOebi

kDrrrr

krr

NCCdtdr

RePrfilm

2 1)(/)(

+−

+

−=−

3 )1( ZXXrr pp +−′=

productreact

reactproduct

MM

Z⋅

⋅=ρρ

′−+

−−=

)1(1

)(

p

ii

g

s

eboMgO

si

rr

rDk

CCN

kdtdr

)

111()1(1

)1)((3

331

32

XZXXXr

Dk

XCCN

kr

dtdX

pg

s

eboMgO

s

p

+−−

−−+

−−

−=rp, k1

rc,k2

c

cs rrfork

rrforkk

<≥

=

2

1

)(0βαXDD gg −=

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 500 1000 1500

MgO

Con

vers

ion

Time (s)

Exp 350 °C

Exp 425°C

Exp 450°C

Model

0

1E-10

2E-10

3E-10

4E-10

5E-10

6E-10

0 0.5 1

Dg

(m2/

s)

MgO Conversion

Model

Parameter

Without Steam With Steam

350 o C 425

o C 450

o C 425

o C

Dg0 1.1E-10 6.0E-10 1.2E-9 2.1E-9

α 3.5E+2 8.1E+1 5.7E+1 4.6E+1

β 3 3 3 3

)( 30 XDD gg α−=

------------------------------------------------------------------------------------------------------------------------------------------- Onischak and Gidaspow, ”Separation of Gaseous Mixtures by Regenerative sorption on Porous Solids. Part II: Regenerative separation of CO2”, Recent Developments in Separation Science, ed. N. Li, 1972

0

5

10

15

20

25

30

0 500 1000 1500 2000 2500 3000 ϕ

(t)

t(s)

Thiele Modulus in our study

45 cm

30 cm

Thermocouple

Thermocouple

8 cm

7 cm

T1

Thermowell

Bed of Sorbent

Quartz Beads

SS Annulus

SS Annulus

Frit

Reactor Body

T2

T1

T2

420

425

430

435

440

445

450

455

0 5 10 15 20 25 30 35 40 45 50Time, min

Tem

pera

ture

, C

T11T12T13T14

420 425 430 435Bed Temp., C

Posi

tion

of th

e be

d, c

m

0

8

4 8 cm

2 cm

T11

T12

T14

T13

Thermowell

420 425 430 435Bed Temp., C

Posi

tion

of th

e be

d, c

m

0

8

4 8 cm

2 cm

T11

T12

T14

T13

Thermowell

@ 425°C start 50 min

CO2 Absorption Breakthrough Curve at Different Operating Temperatures

0

0.5

1

1.5

2

2.5

3

3.5

4

0 5 10 15 20 25 30 35 40Time, min

CO

2 Out

let F

low

, mol

/min

*103

Sorbent: EP68System Pressure= 20 atmInlet Total Flow Rate= 200 cm 3 /minCO2 inlet= 50 %molN2 inlet= 50 %mol

450°C

425°C

350°C

Bed Inlet

Bed Outlet

Results of the CFD model in terms of pressure drops, capturing the cluster formation and CO2 removal rate is in a good agreement with the experimental data

An explicit Reaction kinetics model has been developed which is able to explain TGA experimental data very well and is suitable for CFD applications

PBM and the coupling algorithm for implementation in the CFD code has been developed and verified. More verification is in progress.

Simulations Validation/Verification of the coupled CFD-PBM. Validation of reaction model vs Packed-bed experiments Application of the CFD-PBM in simulation of the circulating fluidized

bed reactor

Experiments Sorbent improvement Reaction rate measurement in shallow/disperse bed reactor

Thanks to the Department of Energy (DOE-NETL) and ICCI for financial support.

38/43

Thanks for your attention

A second order discretization scheme was used to discretize the governing equation domain including 34x1200 uniform rectangular cells.