Benchmarking Mass

Transfer Correlations with a

Nonequilibrium Model
Harry Kooijman & Ross Taylor
Clarkson University
 Packed Column Design
● Column Diameter:
– Capacity: F-factor or C-factor
– Pressure Drop: HETP/∆p
● Column Height H

– Mass Transfer: HETP

Which models to use for

a particular packing? D
 Recent new MTC Models
● Olujic-Delft 1997-2003 [Various]
● Erasmus-Nieuwoudt [IECR, 40, pp.2310-2321]
● Del Carlo-Olujic-Paglianti 2006 [IECR, 45, pp.7967-7976]

How good are these models?

Traditional MTC Model Development
● Total Reflux data (FRI, SRP, TU Delft, Koch-
Glitsch, Sulzer ChemTech, ...)
● Simulation with equilibrium model:
– Determine number of stages
– Plot average HETP versus F or C-factor
– Plot pressure drop versus F or C-factor
● Correlate HETP using:
– One or two-film approach
– Fixed physical properties
 Distillation Test Data
● Typical Systems, Pressures, Thermodynamics:
– c-C6/n-C7, 0.3-4.1 bar, UNIFAC+Antoine
– o/p-Xylene, 0.1-0.3 bar, Ideal+Antoine
– iC4/nC4, 7-11 bar, SRK or PR
– EB/CB, 0.1 bar, Ideal+Antoine
– EB/ST, 0.1 bar, Ideal+Antoine
– MeOH/H2O, 1 bar, NRTL+Antoine

How constant is the HETP?

 HETP vs. Packed Bed Height
HETP varies due to: C hemS ep

10
– T & p changes
20

30

Stage
40

c-C6/n-C7 1atm, 3m bed 50

60

80 85 90 95

T emperature (oC )
 HETP vs. Packed Bed Height
HETP varies due to: C hemS ep

10
– T & p changes
20
– Concentration changes
30

Stage
40

50

60

0 0.2 0.4 0.6 0.8 1

Liquid mole fraction

C 6H 12 C 7H 16
 HETP vs. Packed Bed Height
L iq u id D e n sity (kg /m 3 )

HETP varies due to: 620 640 6 60 680 70 0 720

C h e mS e p

– T & p changes 10

– Concentration changes 20

– Consequent changes in 30

Stage
densities 40

50

60

3 3 .1 3 .2 3 .3 3 .4

Va p o u r D e n sity (kg /m 3 )
 HETP vs. Packed Bed Height
L iq u id Visco sity (N /m 2 .s)

HETP varies due to: 0 5 e -0 0 5 0 .0 0 0 1 0 .0 0 0 1 5

C h e mS e p

– T & p changes 10

– Concentration changes 20

– Consequent changes in 30

Stage
densities, viscosities 40

50

60

7 .6 e -0 0 6 7 .8 e -0 0 6 8 e -0 06 8 .2 e -0 0 6 8 .4 e -0 0 6

Va p o u r Visco sity (N /m 2 .s)

 HETP vs. Packed Bed Height
HETP varies due to: C hemSep

10

– T & p changes
20

– Concentration changes
30

Consequent changes in

Stage
–
densities, viscosities, 40

surface tension 50

60

0 .0 1 3 0 .0 1 4 0 .0 1 5 0 .0 1 6 0 .0 1 7 0 .0 1 8

S u rfa ce te n sio n (N /m )
 HETP vs. Packed Bed Height
HETP varies due to:
10

– T & p changes
20

– Concentration changes
30

Consequent changes in

Stage
–
densities, viscosities,
40

surface tension, and 50

diffusivities 60

0 2 4 6 8 10 12 14

Liquid D iffusivity (1e-8m2/s)

 HETP vs. Packed Bed Height
HETP varies due to:
10

– T & p changes p=1.03bar

20

– Concentration changes
30

Consequent changes in

Stage
–
densities, viscosities, 40
p=0.33bar
surface tension, and 50

diffusivities 60

– And changes in relative 1 .6 1 .6 5 1.7 1 .7 5 1 .8 1 .8 5 1 .9 1 .9 5 2

volatility a lp h a
 HETP vs. Packed Bed Height
HETP varies due to: C hemS ep

10 1.03bar
– T & p changes
20
– Concentration changes 0.33bar
30
– Consequent changes in

Stage
4.13bar
densities, viscosities, 40

surface tension, and 50

diffusivities
60
– And changes in relative
volatility 0 .2 5 0.3

HET P
0 .3 5
 HETP vs. Packed Bed Height
HETP varies due to: C hemS ep

10 1.03bar
– T & p changes
20
– Concentration changes 0.33bar
30
– Consequent changes in

Stage
4.13bar
densities, viscosities, 40

surface tension, and 50

diffusivities
60
– And changes in relative
volatility 0 .2 5 0.3

HET P
0 .3 5

We must average HETP

over the bed height!
 A Different Approach
● Data for multiple systems/pressures
● Simulate in nonequilibrium model (ChemSep)
● Compute HETP from back-calculated efficiency
● Average HETP over the whole packed bed
Problem: Often no concentration gradient
published. Use educated guess from T & p
 Collecting Data
● “ScanIt”
 Collecting Data
● “ScanIt”
● Simulate it:
ChemSep
Total Reflux
 Collecting Data
● “ScanIt”
● Simulate it:
ChemSep
Total Reflux
● Parametric
Study to plot
average HETP
vs F-factor
Collecting Data
B X o /p-Xyle ne 1 6 T orr @ F R I (1 .2 2m ID )

0 .3

0.2 5

0 .2

Average HETP (m)

0.1 5

0 .1

0.0 5

0
0 1 2 3 4

F -facto r

BRF85 exp.
 Model Fitness
B X o /p -Xyle n e 1 6 T o rr @ F R I (1 .2 2 m ID )

0 .3

0 .2 5

0 .2
Average HETP (m)

0 .1 5

0 .1

0 .0 5

average error = 35mm

0
0 1 2 3 4

F -fa cto r

BZ95 e xp . E rro rs
 Structured Packing Test Data
● Sulzer Mellapak 250Y, Mellapak Plus 252Y
● Montz B1-250, B1-250M
● Koch-Glitsch Flexipac 2Y, Flexipac HC
● Raschig SuperPak 300
● Sulzer BX, BX-Plus
 MTC Models
Gauze Metal Structured Packing:
● Zogg(+Toor-Marchello) 1983 [Chem.Ing.Tech., 45, p.67]
● Bravo-Fair 1985 [Hydrocarbon Processing, January]
● Brunazzi 1995 [Chem.Eng.Technol., 19, pp.20-27]
● Bravo-Rocha-Fair 1996 [IECR]
 MTC Models
Sheet Metal Structured Packings:
● Bravo-Rocha-Fair 1992/1996 [DA1992, IECR]
● Billet-Schultes 1992 [Chem.Eng.Technol., 16, pp.370-375]
● Ronge 1995 [PhD]
● Olujic-Delft 1997-2003 [various]
● Erasmus-Nieuwoudt [IECR, 40, pp.2310-2321]
● Del Carlo-Olujic-Paglianti 2006 [IECR, 45, pp.7967-7976]
 Sulzer-BX
B X o /p -Xyle n e 1 6 T o rr @ F R I (1 .2 2 m ID )

0 .3

0 .2 5

0 .2
Average HETP (m)

0 .1 5

0 .1

0 .0 5

0
0 1 2 3 4

F -fa cto r

BZ95 e xp . ZT M 83 BR F96 BR F85

 Sulzer-BX
B X o /p -Xyle n e 7 3 0 T o rr @ F R I (1 .2 2 m ID )

0 .3

0 .2 5

0 .2
Average HETP (m)

0 .1 5

0 .1

0 .0 5

0
0 1 2 3

F -fa cto r

BR F85 e xp . BZ95
 Sulzer Mellapak 250Y
M 2 5 0 Y C B /E B 1 0 0 mb a r @ S u lze r (1 ID )

0 .6

0 .5

0 .4
Average HETP (m)

0 .3

0 .2

0 .1

0
0 0 .0 2 0 .0 4 0 .0 6 0 .0 8 0 .1 0 .1 2

C -fa cto r

BR F92 E xp . S u lze r
 Montz B1-250M
B 1 -2 5 0 M cC 6 /n C 7 1 .0 3 b a r

0 .6

0 .5

0 .4
Average HETP (m)

0 .3

0 .2

0 .1

0
0 0 .5 1 1 .5 2 2 .5 3 3 .5

F -fa cto r

BS92 BR F85 BRF96

E xp . H E T P BR F92 R95
Billet-Schultes MTC: Effect CV
B 1 -2 5 0 M cC 6 /n C 7 1 .0 3 b a r

0 .6

0 .5

0 .4 0.2
Average HETP (m)

0 .3 0.3

0 .2
0.4

0 .1

0
0 0 .5 1 1 .5 2 2 .5 3 3 .5

F -fa cto r

B S '9 2 C v= 0 .3 C l= 0 .9 B S '9 2 C v= 0 .2
E xp . H E T P B S '9 2 C v= 0 .4
Billet-Schultes MTC: Effect CL
B 1 -2 5 0 M cC 6 /n C 7 1 .0 3 b a r

0 .6

0 .5

0 .4
0.3
Average HETP (m)

0 .3 0.5
0.9
10
0 .2

0 .1

0
0 0 .5 1 1 .5 2 2 .5 3 3 .5

F -fa cto r

B S '9 2 C v= 0 .3 C l= 0 .9 B S '9 2 C v= 0 .3 C l= 0 .5
E xp . H E T P B S '9 2 C v= 0 .3 C l= 1 0 .0
B S '9 2 C v= 0 .3 C l= 0 .3
Koch-Glitsch Flexipac 2
F le xip a c 2 cC 6 /n C 7 0 .3 3 b a r

0 .6

0 .5

0 .4
Average HETP (m)

0 .3

0 .2

0 .1

0
0 0 .5 1 1 .5 2 2 .5 3 3 .5

F -fa cto r

R 95 E xp . H E T P
 Raschig SuperPak 300
S u p e rP a k 3 0 0 cC 6 /n C 7 1 .6 5 b a r

0 .6

0 .5

0 .4
Average HETP (m)

0 .3

0 .2

0 .1

0
0 0 .5 1 1 .5 2 2 .5 3 3 .5

F -fa cto r

BR F92 E xp . H E T P

Lack of geometry data (other than Ap): Estimated deq 30mm

 Conclusions
● Consistent HETP comparisons for c-C6/n-C7
● Public distillation test data collection
(work in progress)
● Overall best MTC correlations (so far):
– Gauze packings: Brunazzi '95
– Sheet metal packings: Bravo-Rocha-Fair '92
– New models do not provide better predictions
 Future Work
● Compare pressure drop & capacity models
● Benchmark random packing models
● Model liquid entrainment @ flood in MTC?
Questions?