chemical reaction engineering laboratory data base expansion for bubble column flows data base...

CHEMICAL REACTION ENGINEERING LABORATORY

Data Base Expansion for Bubble Data Base Expansion for Bubble Column FlowsColumn Flows

reported for DOE in period from 1995 to 2001

& summarized by Peter Spicka


OutlineOutlineTopical review of the data reported for DOE since 1995• Gas Holdup and Liquid Recirculation• Solid Loadings and Sparger Effect• Scale-Up of Bubble Columns• Eddy Diffusivities• Summary

GoalsGoalsContract DE-FC 22-95 PC 95051Contract DE-FC 22-95 PC 95051• Development of reliable data base for evaluation of parameters in CFD

based models• Development of improved engineering models for flow mixing and mass

transfer in bubble columns


Gas holdup and Liquid RecirculationGas holdup and Liquid Recirculation

Commonly used correlationsCommonly used correlations

009.0296 16.019.098.044.0 GLgU

17.027.0

3

487.0

4.01

L

G

L

LLg gU

nRrccn

nr

122

2

31,0 4.1 BgCL UUDU

310 737.0 CgL DUU

NN

L

L

R

r

U

rU

2

0

21

Reference Axial Liquid Velocity

Overall gas holdup

Reilly et al. (1986)

Hammer et al. (1984)

Gas holdup radial profile

Luo & Svendsen (1991)

Centerline axial liquid velocity

Joshi & Sharma (1979)

Zehner (1982b)

Axial liquid velocity profile

Garcia-Calvo et al. (1994)

Hydrodynamics is driven by:• buoyancy, drag, inertia, pressure,

viscous, interface forces… • strong coupling between the forces• many different scales

The scale-up is tricky and only very sophisticated CFD simulations can resolve all the aspects - feasible in future ?

Main aspects of multiphase flow

Simplifications:

• Steady–state one-dimensional flow

• Only gas holdup, liquid velocity and turbulence radial profiles are determining factors


Gas Holdup & Liquid Recirculation I.Gas Holdup & Liquid Recirculation I. Effect of UEffect of Ugg and Liquid Properties and Liquid Properties (DOE Quarterly Reports 7-11, 1996)(DOE Quarterly Reports 7-11, 1996)

-60.0

-40.0

-20.0

0.0

20.0

40.0

60.0

0.0 0.2 0.4 0.6 0.8 1.0r/R

Axi

ally

Ave

rag

ed

Axi

al V

elo

city

, cm

/s

Air-Water

Air-Drakeoil

Ug=10 cm/s

Gas holdup profiles

Air-water air-Drakeoil

Observations:• Increased Ug results in higher holdup and less uniform holdup profile • Gas holdup is lower in air-Drakeoil system due to higher liquid viscosity (20 cP)• The higher holdup in air-water system results in higher recirculation rate

Axial velocity profiles

air-water air-water vs. air-Drakeoil

air-water

air-Drakeoil


Gas Holdup & Liquid RecirculationGas Holdup & Liquid Recirculation II. II.Effect of Column Diameter & Internals Effect of Column Diameter & Internals (DOE Quarterly Reports 8 & 9, 1996)(DOE Quarterly Reports 8 & 9, 1996)

-30.0

-20.0

-10.0

0.0

10.0

20.0

30.0

40.0

50.0

0.0 0.2 0.4 0.6 0.8 1.0r/R

Tim

e A

vera

ged L

iquid

Axi

al V

elo

city

, cm

/s Without internals

With internals

Ug=10 cm/s

Overall gas holdup in 6”, 8” and 18” columns

18”8”6”

Internals layout

Axial velocity profiles

6”& 8” columns with and without internals in 18” columns

Observations:• Overall gas holdup and liquid recirculation increases with column diameter • Effect of internals on axial velocity is less pronounced• Internals reduce radial eddy diffusivity ( Chen et al., 1999)


Gas Holdup & Liquid RecirculationGas Holdup & Liquid Recirculation III.III.Gas Distributor Effect Gas Distributor Effect (DOE Quarterly Report 18, 1999)(DOE Quarterly Report 18, 1999)

Gas distributors:D1D1

Porosity = 0.1 %163 holes of 0.4 mm ID

Equilateral triangle 1 cm apart

D2D2


Distributed on a cross

Porosity = 0.1 %Single hole of 5.1 mm ID

Located in the center

D3D3D4D4


Equilateral triangle 1 cm apart

Porosity = 0.04 %61 holes of 0.4 mm IDCircular rings 1.5 cm

apart

D5D5D6D6

Porosity = 1.0 %163 holes of 1.25 mm IDEquilateral triangle 1 cm

apart

0

0.1

0.2

0.3

0.4

0 0.2 0.4 0.6 0.8 1

Dimensionless radius

Gas

Hol

dup

D1D2D3D4D51.05 Average0.95 Average

z/D = 9.0Ug = 14 cm/s

0

0.1

0.2

0.3

0.4

0.5

0 0.2 0.4 0.6 0.8 1

Dimensionless radius, r/R

Gas

Hol

dup

D1D2D3D41.05 Average0.95 Average

z/D = 9.0Ug = 30 cm/s

0

0.1

0.2

0.3

0.4

0.5

0 0.2 0.4 0.6 0.8 1


Gas

Hol

dup

D1D2D3D4D51.05 Average0.95 Average

z/D = 2.1Ug = 14 cm/s

0

0.1

0.2

0.3

0.4

0.5

0 0.2 0.4 0.6 0.8 1

Dimensionless radius

Gas

Hol

dup

D1D2D3D41.05 Average0.95 Average

z/D = 2.1Ug = 30 cm/s

Ug = 14 cm/s Ug = 30 cm/s

Z+

Z-

Observations:• Gas distributor effect is visible only at low Ug (14 cm/s) and near the

column bottom• Flow stabilizes faster when single nozzle distributors are used• At Ug of 30 cm/s, the gas distributor effect is negligible


Gas Holdup & Liquid RecirculationGas Holdup & Liquid Recirculation III. III.Gas Distributor Effect Gas Distributor Effect (DOE Quarterly Report 14, 1996; Degaleesan (DOE Quarterly Report 14, 1996; Degaleesan et al.,et al., 1997) 1997)

Liquid Velocity profile

8” column, Ug=12.0 cm/s. Distributors are: Cone (8C), Bubble Cap (8B), and Perforated Plate (8A)

Turbulent kinetic energy profile

Observations:Single nozzle distributors • produce larger bubbles flow is less organized with large spiraling structures• suppressed recirculation and higher turbulent kinetic energy (about 40% higher compared to multiple nozzles

distributors)

Multiple holes distributors • Smaller bubbles, less violent flow and enhanced recirculation


Gas Holdup & Liquid RecirculationGas Holdup & Liquid Recirculation IV. IV.Pressure Effect Pressure Effect (DOE Quarterly Report 22, 2000; Kemoun (DOE Quarterly Report 22, 2000; Kemoun et al.,et al., 2001) 2001)

• 6.4” column, axial level z/D = 5.5, distributor D4

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.2 0.4 0.6 0.8 1


Ga

s H

old

up

P = 1 atm

P = 4 atm

P = 10 atm

P = 1 atm

P = 4 atm

P = 10 atm

P = 1 atm

P = 4 atm

P = 10 atm

z/D = 5.5

14

Ug cm/s

8

2

Findings• Almost uniform gas holdup profiles,

which are not pressure dependent were observed at lower Ug. This finding is in good agreement with Letzel (1997)

• In churn-turbulent regime, gas holdup increases with pressure (in the interval from 1 to10 bars) as well as the steepness of the profile

• Single hole distributor D3 provides visibly higher gas holdup than perforated plate distributors at 4 bars and Ug of 30 cm/s due to increased dispersion and break up of the gas jet produced by the single nozzle (Kling, 1962; Nauze et al., 1974)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.2 0.4 0.6 0.8 1

Dimensionless Radius, r/R

Ga

s H

old

up

D1D2D3D4D60.95*Avg1.05*Avg

6.4” column, 4 bars, Ug= 30 cm/s

6.4” column


Gas Holdup & Liquid RecirculationGas Holdup & Liquid Recirculation V. V.Effect of Solids Loading Effect of Solids Loading (DOE Quarterly Report 12, 1998)(DOE Quarterly Report 12, 1998)

Influence of Solids Loading Axial Distance

Axial Velocity Profiles

-30

-20

-10

0

10

20

30

40

0.0 0.2 0.4 0.6 0.8 1.0r/R

Ua

x, c

m/s

x=5.6 cm

x=11.9 cm

x=36.9 cm

x=61.9 cm

x=86.9 cm

x=111.9 cm

pp-u8-gls

b)

Dc=4in.Axial Velocity Profiles

-30

-20

-10

0

10

20

30

40

50

0.0 0.2 0.4 0.6 0.8 1.0r/R

Ua

x, c

m/s

u9.6

u8-wt7

u8-wt14

u8-wt20

Z=36.87cmDc=6in.

a)

G-L system G-L-S system

Dc inch(cm) 4 (10.2) 6 (14) 4 (10.2) 6 (14)

Composition air - 50% iso-propanol

air - water air - 50% iso-propanol - alumina

air - water - glass beads

Ug [cm/s] 4-12 2.4-12 2-8 2-14

Solids [wt. %] ‑ ‑ 10 7, 14 and 20

Particle size [m]

- - 40-106 125-177

Sparger perf.plate,bubble cap

perf. plate sintered plateperforated plate

perforated plate

Comparison of G-L and G-L-S systems

•Ug has smaller effect on axial velocity profiles in slurries and its effect decreases with increased concentration of slurries

•All observed differences can be attributed to altered viscosity and density of the pseudo slurry phase


Eddy DiffusivitiesEddy DiffusivitiesEffect of Effect of UgUg, solids loading , solids loading (DOE Quarterly Reports 13 & 14, 1998)(DOE Quarterly Reports 13 & 14, 1998)

Axial Eddy Diffusivity

0

50

100

150

200

250

300

0.0 0.2 0.4 0.6 0.8 1.0

r/R

Dzz

, cm

^2/s

u12

u14-wt7

u14-wt14

u14-wt20

Axial Eddy Diffusivity

0

50

100

150

200

250

300

0.0 0.2 0.4 0.6 0.8 1.0

r/R

Dzz

, cm

^2/s

u2-wt7

u8-wt7

u14-wt7

Effect of Ug Effect of solids loading

Eddy diffusivity • Flow in bubble columns is of transient nature• Fluctuating character of flow and backmixing

can by captured by eddy diffusivity, defined in Langrangian framework as:

ijjijiij yvyvdt

dyy

dt

dtD

2

1

2

1

6” column , GLS system, glass particles of 150 m

Findings of the study• With increased Ug, the axial eddy

diffusivity increases• Maximum Dzz occurs at r/R = 0.75• Solids loading does not affect the axial

eddy diffusivity profiles significantly


Eddy DiffusivitiesEddy DiffusivitiesEffect of Effect of UgUg and and DcDc – – correlations correlations (DOE Quarterly Report 13, 1998)(DOE Quarterly Report 13, 1998)

3.03.08.0

2 6.1062325

gCC

zz UDD

scmD 3.03.08.0

2 0.13350

gCC

rr UDD

scmD

These correlation are valid for air-water systems in large columns size (Dc> 10 cm) and churn-turbulent regime ( Ug > 5 cm/s)


Scale–Up Issues I.Scale–Up Issues I.((DOE Quarterly Report 13, 1998)DOE Quarterly Report 13, 1998)

CREL contribution Ibased on CARPT/CT, a new correlation for centerline axial velocity was developed:

nRrccn

nr

122

2

004.0146.0598.03 Re10188.2 LGG MoFrn2492.02 Re1032.4 Gc

44.065.2

44.0

0

65.21cn

L

L

R

rcn

U

rU

Wu et al. (2001a) Wu et al. (2001b)

Chronology:Energy balance models:• Whalley & Daviddson (1974)• extended by Joshi & Sharma (1979)

who proposed circulation structuresMomentum balance models:• Rietema & Ottengraf (1970)• Zehner (1980)Implementation of turbulence:• universal mixing length by Ueyama &

Miyauchi (1979)• Anderson and Rice (1989) proposed

a‘three zones’ concept• Geary and Rice (1992) proposed

model which depends on bubble size

4.04.02.2/ gCC UDscmU

CREL contribution IIIn a separated effort ( not directly funded by DOE), new correlations for radial gas holdup and axial liquid velocity were proposed


Scale –Up Issues II.Scale –Up Issues II.((DOE Quarterly Report 13, 1998: Degaleesan, 1997)DOE Quarterly Report 13, 1998: Degaleesan, 1997)

1653.00717.50929.5 2

2

2

Pwhere

PDD rrrr 5847.0005035.02404.34979.3 34

4

4

Pwhere

PDD zzzz

CREL contribution IIICorrelations for radial profiles of axial and radial eddy diffusivities were proposed

(Valid for churn-turbulent regime and Dc> 10 cm)


Summary I.Summary I.

Extensive data base has been created which Extensive data base has been created which encompass broad range of operating conditions:encompass broad range of operating conditions:

• Column diameters: 4, 6, 8 and 18 inch column

• Range of UG : from 2 to 30 cm/s

• Range of pressure: 1, 4, and 10 bars for 6.4” column• Distributors: perfor. plate (various porosities), sintered

plate,cross sparger, cone, buble cap • Liquids : water, Drakeoil, 50% isopropanol in water• Internals in 18”column)


Summary II.Summary II.• Gas holdup and liquid recirculation is affected primarily by superficial gas

velocity• Secondary effects are due to: liquid physical properties; column diameter and;

solids concentration• Heat exchange tubes do not affect significantly gas holdup and liquid

recirculation profiles• Gas distributor does has minimum effect gas holdup and liquid velocity in fully

developed region• Gas holdup is not affected by pressure in bubbly regime but it rises with

pressure in churn-turbulent regime and becomes increasing parabolic• Correlations have been developed for the gas holdup, liquid velocity and eddy

diffusivity radial profiles

Future workFuture work

• Data base extension to higher pressure and temperature• Experiments at low H/Dc to identify sparger with most desirable properties• New models need to be proposed as to which variable is dominant one that

governs the establishment of gas holdup and liquid velocity profile

chemical reaction engineering laboratory data base expansion for bubble column flows data base...

Documents

higher holdup

gas distributors

higher liquid viscosity

airdrakeoil system

airwater system results

apart porosity

id equilateral triangle

cms z z zz