performance of gas-liquid reactors multiphase flows to …mjm/kansas_slides.final.pdf · • Gas-liquid packed bed reactors are ... • Generic reactor system for gas-liquid w/ solid

[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept

Use of hydrodynamic understanding ofmultiphase flows to improve

performance of gas-liquid reactors

Mark J. McCready

University of Notre Dame


Acknowledgements

• Graduate students:– Ben Wilhite, Richard Hwang

• Undergraduates”– James Kacmar, Brandon Blackwell

• Faculty colleague:– Arvind Varma


Background• Gas-liquid packed bed reactors are

currently used for hydrotreating ofpetroleum, some waste treatment andother oxidation reactions

• Generic reactor system for gas-liquid w/solid catalyst that does not have real highheat removal requirements– heat removal is one problem in hydrotreating

leading to “hot spots” and “coke balls”


Background (cont.)

• If selectivity can be improved andhydrodynamics better understood weenvision these as a useful reactor toproduce small quantities of hazardouschemicals “on location”, (closet scaleproduction)


Objectives of our work• Develop detailed knowledge about the

hydrodynamics of gas-liquid flows and useit to understand and design multiphasereactors that take advantage of thespecific dynamics of these flows

• Would like these reactors to be optimizedas “smallest possible volume”, “highestreasonable selectivity”,

• Want: “No surprises”.


Why do we think there is potential forsomething new and exciting here?

• The dynamics of these flows cause strongvariations in pressure drop and heat andmass transfer

• These have not been understood orexploited to enhance reaction


Preview• We will show the interesting dynamics of a gas-

liquid packed bed flow– Time varying behavior, strong pressure fluctuations

• Pulses grow with distance like a convectiveinstability– This allows us to do reaction studies both with and without

pulses present

• Pulses significantly affect the reaction behavior• Detailed studies of local heat transfer show

– Most of the heat removal occurs during a pulse

– The rate of heat removal in a pulse is about the same asonly a liquid flow

– Pressure drop and heat transfer do not scale exactly thesame leading to some degree of possible optimization


Trickle flow


Intermittent pulsing


Pulsing


16.216.015.815.615.415.2P

(ps

ia)

6050403020100time (sec)

G=0.345 kg/m2-s

16.015.815.615.415.215.014.8P

(ps

ia)

6050403020100time (sec)

G=0.33 kg/m2-s L = 7.9 Kg/(m2 -s)

16.216.015.815.615.415.2P

(ps

ia)

6050403020100time (sec)

G=0.40 kg/m2-s

16.416.216.015.815.615.4P

(ps

ia)

6050403020100time (sec)

G=0.52 kg/m2-s

Pulse occurrence at increasing G


Different flow regimes

Stratified

Slug (pulse)


Surprising result: Pressure drop ratioUsing a simple model, air-water flow in a small channel


Effect of hydrodynamics

• We expect that the time-varyinghydrodynamics plays a direct role on thereaction through the fluctuating heat andmass transfer rates.

• First check this with modeling


CSTR model with fluctuatingmass transfer

R. Wu, M. J. McCready and A. Varma (1995), Chemical Engineering Science, 50, pp3333-3334.


Mass transfer fluctuations


Calculated enhancement by pulsingSignificant enhancement, interesting frequency effect

R. Wu, M. J. McCready and A. Varma (1995), Chemical Engineering Science, 50, pp3333-3334.

http:// www.nd.edu/~mjm/Reaction_Selectivity_Mult.nb

Frequency isdimensionless with first orderrate constant


How can we do an experiment ?• Pulses behave as a “convective instability”

which means they gain strength withdistance in the flow direction.

• Thus you can find ranges of flowrateswhere pulses grow slow enough that theinception region is about 1/2 of the column

• At the top: no pulses• At the bottom: pulses


Wave measurements as function of distance showingthe development of large disturbances

10-6

10-5

10-4

10-3

10-2

wave spectrum (cm2-s)

0.12 3 4 5 6 7

12 3 4 5 6 7

102 3 4 5 6 7

100frequency (1/s)

40

30

20

10

0

-10

growt

h rat

e (1/

s)

RL = 300RG = 9975

µL = 5 cP

linear growth k-ε model

wave spectrum @ 1.2 m 3.8 m 6 m

Data of Bruno andMcCready, 1988


Location of the first appearance ofpulsing

50 100 150 200 250 30050

100

150

200

250

300

350

400

450

500

550

Rel

puls

e he

ight

/ v l

Reg = 187

Reg = 375

Reg = 600

Reg = 750

Reg = 2250


Catalyst configurationConfiguration Configuration

25 cm

25 cm

8 cm

8 cm

5 cm

5 cm

catalyst

TricklingRegime

PulsingRegime

Transition

"Lower Packing" "Upper Packing"


Model reaction

+2H2+H2

+H2

R1

R2R3

Phenylacetylene (PA) Styrene (ST)

Ethylbenzene (EB)

C C C C

C C


Reaction experiment results

• Pulses atbottom,no pulsesat top

0 10 20 30 400

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time, Minutes

Dim

ensio

nle

ss C

oncentr

ation, C

x / C

PA

,o

A1: Pulsing A2: Trickling

PA

ST

EB


Reaction results

0 10 20 30 400

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time, Minutes

Dim

en

sio

nle

ss C

on

ce

ntr

atio

n,

Cx /

CP

A,o

B1: Pulsing, LPB2: Pulsing, UP

PA

ST

EB

• Pulses atbottom,and top


More reaction results

0 10 20 30 400

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time, Minutes

Dim

ensio

nle

ss C

oncentr

ation, C

x / C

PA

,o

C1: Pulsing C2: TricklingPA

ST

EB


Reaction model• CSTR with reactor and catalyst

Reactor:

Vr �dCPAr

d t

= Kc;PA � A � (CPAps

� CPAr)

Vr �dCSTr

d t

= Kc;ST � A � (CSTps � CSTr)

Vr �dCEBr

d t

= Kc;EB � A � (CEBps

� CEBr)

Pellet:

"p �@ CPAp

@ t

= De;PA �

@2 CPAp

@ r2

� (k1 � CPAp

+ k3 �CPAp

)

"p �@ CSTp

@ t

= De;ST �

@2 CSTp

@ r2

+ (k1 � CPAp

� k2 � CSTp)

"p �@ CEBp

@ t

= De;EB �

@2 CEBp

@ r2

+ (k2 � CSTp + k3 � CPAp)


Detailed modeling results


Effect of pulses on reaction• We can observe an effect of pulses on selectivity

-- 40% max effect for our system• Reaction modeling suggests that time-varying mass

transfer is the reason• To further test and fully exploit we need to be

able to control strength and frequency of pulses• We to understand the mechanism of formation!


How do we think about the flow in the column?

• Are the gas and liquid homogenouslydispersed?

• Is the other limit where there are liquid-rich and gas rich regions?


Heterogeneous basestate ?

Liquid Liquid

Gas

liquid rich regions

liquid rich regions


Pulsing in a horizontal column

2.82.42.01.6P

ress

ure

(Psi

)

3025201510Time (sec)

L = 15.5 Kg/m2 sG = 0.49 Kg/m2 s

3.22.82.42.0P

ress

ure

(Psi

)

3025201510Time (sec)

L = 15.5 Kg/m2 sG = 0.58 Kg/m2 s

3.22.82.4P

ress

ure

(Psi

)

3025201510Time (sec)

L = 15.5 Kg/m2 sG = 0.66 Kg/m2

s

2.42.01.61.2P

ress

ure

(Psi

)

3025201510Time (sec)

L = 15.5 Kg/m2 sG = 0.41 Kg/m2 s

Pressure tracings as a function of gas velocity ranging from mild waves to pulses

Disturbed

Disturbed

Pulsing

Heavy Pulsing


Detailed measurements of pulsing invertical column, local heat transfer


Localinstantaneousheat transfercompared to

pressuredrop

5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 100

0.5

1

1.5

2

2.5

3

3.5

x 104

time, second

h t, w/(

K ⋅

m2 )

5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 100

50

100

150

200

250

300

time, second

Ts, o C

5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 100

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

time, second

P, p

sig


Heat transfer enhancement by pulsing• Ratio of pulse/base heat transfer rate

0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.20.02

0.03

0.04

0.05

0.06

0.07

vg, m/sec

v l, m/s

ec

h

1.51.522 2.52.5

2.5

3

3

33.5

3.5

3.5

3.5

4

4

44

4

4.54.5


Effect of gas and liquid on heat transfer

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.080

0.5

1

1.5

2

2.5

3x 10

4

vl, m/s

h t,p, w

/(K

⋅ m

2 )

vg=0.32 m/s

vg=0.65 m/s

vg=0.97 m/s

vg=1.29 m/s

vg=0 m/s

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.32000

4000

6000

8000

10000

12000

14000

vg, m/s

h t,b, w

/(K

⋅ m

2 )

vl=0.018 m/s

vl=0.036 m/s

vl=0.054 m/s

vl=0.072 m/s

vl=0 m/s


Can heat transfer be optimized?

0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.20.02

0.03

0.04

0.05

0.06

0.07

vg, m/s

v l, m/s

ht,avg

, w/(K ⋅ m2)

50005000 1000

1000010000

15000

1500015000

15000

20000

2000020000

20000

2500025000

0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.20.02

0.03

0.04

0.05

0.06

0.07

vg, m/s

v l, m/s

Pavg

, psig

0.2 0.4

0.4

0.4

0.6

0.6

0.6

0.6

0.8

0.8

0.8

0.8

1

1

1

1.2

Heattransfer

Pressuredrop


Ratio of thermal energy removedto mechanical energy expended


Ratio of thermal energy removedto mechanical energy expended


Continuing work

• Hydrodynamics– “2D” packed bed

• Reaction– Catalysts that have pore sizes that enable

convective transport within the pellets caused bypressure fluctuations from the pulses.


Configuration of two-dimensional bedsGas Liquid

Gas Liquid

Gas Liquid


Pressure tracing in 2D bed


Pressure tracing in 2D bed


Catalyst structure

Macro channelsspan catalyst pellet

Nanoporous structure with macro channels.

blow up of region


Conclusions• Complex time-varying hydrodynamics occur

in gas-liquid packed bed systems– These occur in pipe flow as well.

– Flow regime significantly affects the pressure dropand transport rates

• Because pulses behave as a convectiveinstability, it is possible to do experimentswhere the flow rates are the same forpulsing and non-pulsing flows


Conclusions (cont.)• These experiments show that pulses definitely

affect the selectivity of a sequential reaction.– Modeling suggests this is because of the fluctuating mass

transfer coefficient having a time scale that allows interactionbetween mass transfer and reaction

• Detailed measurements of the local instantaneousheat transfer suggest– Most of the heat removal occurs during a pulse

– The rate of heat removal in a pulse is about the same asonly a liquid flow

– Pressure drop and heat transfer do not scale exactly thesame leading to some degree of possible optimization


Dimensionless groups are ratios ofimportant effects in a problem

Re

≡ Inertia ForcesViscous Forces


Dimensionless groups do not needto be on technical subjects

CrHow Smart You Are

How Smart You Think You Are≡


DimensionlessConfucius Proverb

• He who knows not and knows he knows notis a child, teach him, Cr~1

• He who knows not and knows not he knowsnot is a fool, shun him, Cr<<1

• He who knows and knows not he knows isasleep, awaken him, Cr>>1

• He who knows and knows he knows is wise,follow him Cr~1

Documents

performance of gas-liquid reactors multiphase flows to …mjm/kansas_slides.final.pdf · • Gas-liquid packed bed reactors are ... • Generic reactor system for gas-liquid w/ solid