performance of gas-liquid reactors multiphase flows to …mjm/kansas_slides.final.pdf · •...
TRANSCRIPT
[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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Acknowledgements
• Graduate students:– Ben Wilhite, Richard Hwang
• Undergraduates”– James Kacmar, Brandon Blackwell
• Faculty colleague:– Arvind Varma
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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”
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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)
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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”.
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Different flow regimes
Stratified
Slug (pulse)
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Surprising result: Pressure drop ratioUsing a simple model, air-water flow in a small channel
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
CSTR model with fluctuatingmass transfer
R. Wu, M. J. McCready and A. Varma (1995), Chemical Engineering Science, 50, pp3333-3334.
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Catalyst configurationConfiguration Configuration
25 cm
25 cm
8 cm
8 cm
5 cm
5 cm
catalyst
TricklingRegime
PulsingRegime
Transition
"Lower Packing" "Upper Packing"
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Model reaction
+2H2+H2
+H2
R1
R2R3
Phenylacetylene (PA) Styrene (ST)
Ethylbenzene (EB)
C C C C
C C
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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)
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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!
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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?
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Heterogeneous basestate ?
Liquid Liquid
Gas
liquid rich regions
liquid rich regions
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Detailed measurements of pulsing invertical column, local heat transfer
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Ratio of thermal energy removedto mechanical energy expended
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Ratio of thermal energy removedto mechanical energy expended
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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.
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Configuration of two-dimensional bedsGas Liquid
Gas Liquid
Gas Liquid
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Catalyst structure
Macro channelsspan catalyst pellet
Nanoporous structure with macro channels.
blow up of region
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Dimensionless groups are ratios ofimportant effects in a problem
Re
≡ Inertia ForcesViscous Forces
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
Dimensionless groups do not needto be on technical subjects
CrHow Smart You Are
How Smart You Think You Are≡
[email protected] http://www.nd.edu/~mjm http://www.nd.edu/~chegdept
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