Two-Phase Gas/Liquid Pipe Flow: Ron Darby PHD, Pe Professor Emeritus, Chemical Engineering Texas A&M University
Two-Phase Gas/Liquid Pipe Flow: Ron Darby PHD, Pe Professor Emeritus, Chemical Engineering Texas A&M University
»Solid-Gas
»Solid-Liquid
»Gas-Liquid
»Liquid-Liquid
Gas-Liquid Flow Regimes
Homogeneous
Highly Mixed
“Pseudo Single-Phase”
High Reynolds Number
1/ 2
W 2
L
W
L W L
Vertical Pipe Flow Regime Map
DEFINITIONS
m m L mG LQL G QG
Mass Flux (G):
m m L mG
G GL GG
A A A
Volume Flux (J):
G GL GG
J J L JG
m L G
QL QG
Vm
A
Volume Fraction Gas: ε
Vol. Fraction Liquid: 1-ε
Phase Velocity:
JL JG
VL , VG
1
Slip Ratio (S): Mass Fraction Gas (Quality x):
VG mG
S x
VL mG m L
mG x G
S
mL 1 - x L 1 -
Density of Two-Phase Mixture:
m G 1 L
where
x
ε
x S 1 x ρG / ρL
is the volume fraction of gas in the mixture
Holdup (Volume Fraction Liquid):
S 1 x ρG / ρL
φ 1 ε
x S 1 x ρG / ρL
Slip (S)
1 K 1 x / x ρG / ρL
S
K 1 x / x ρG / ρL
where
K 1 0.12 / Z
19
0.95
/ 1 ε
1/ 4
ZN 1/ 6
Re N 1/ 8
Fr
HOMOGENEOUS GAS-LIQUID PIPE
FLOW
Energy Balance (Eng’q Bernoulli Eqn)
2
2 f mG dx dz
G νGL
2
ρm g
dP ρm D dL dL
dL 2 dνG
1 xG
dP
where
ν GL ν G ν L 1 / ρG 1 / ρ L
For “frozen” flow (no phase change):
dx
0
dL
νG νG
1/ k
1 P
P ρP 1
, 1 k / k
T P s ρ1 kP
For frozen ideal gas/liquid choked flow:
ρm kP
G cρm
ε
νG ν cT 2
P λ
GL p
T
2
GL
Homogeneous Horizontal Flow
2 fm
2
G dL νGLdx ρm dz
ρm D
dP
1 xG dνG / dP
2
c p To Ts
x
λGL
Finite Difference Solution – solve for L
D
ΔL -
4 fm
ΔP G 2
Δν gΔZ / ν 2
νG 2
ΣK fit
Dimensionless
4 f m ΔL 2
ΣK fit - Δη G Δε gΔZ / Po νo ε *2
*2
D εG
where
Δη ΔP / Po
G* G / Po ρo G νo / Po
ε ν / νo ρo / ρ
L= ΔL i
i
the pipe inclination angle
Using Z Lcos , where θ is
with the vertical
4 f m ΔL
2 Δη G *2
Δε / εG *2
ΣK
fit
-
D 1 gD cos θ / 4 fPo νo ε
Procedure: Find G, Given L, Po and Pe
P 2 P
L ΦR L
fm fR
Reference Single-Phase Flow:
P 2 P
L ΦLm L
f fLm
or:
P 2 P
L ΦGm L
f fGm
L-M Two-Phase Multipliers
C 1
Φ 2
Lm 1 2
χ χ
Φ 2
Gm 1 Cχ χ 2
State Liquid Gas C
tt turbulent turbulent 20
vt laminar turbulent 12
tv turbulent laminar 10
vv laminar laminar 5
L-M Correlating Parameter
P P
χ
2
L f Lm L fGm
2 f Lm 1 x G
2
P
2
L fLm ρL D
P
2 2
2 fGm x G
L fGm ρG D
Friction Factors
1 x GD
N ReLm
μL
f Gm is based on “gas-only” Reynolds No.
xGD
N ReGm =
mG
Duckler et al. (1964)
ρL
Φ 2
Lm α φ β φ
ρ
and
ρG
Φ 2
Gm α φ β φ
ρ
where
φ 1 ε , ρ are “no slip” values
and
lnφ
α φ 1
1.281 0.478 lnφ 0.444 lnφ
2
ρG 1 φ
2
ρL φ 2
β φ
ρ φm ρ 1 φm
The Reynolds Number is based on mixture
properties:
DG
N Rem β φ
μm
Sizing Relief Valves
for Two-Phase Flow
m
A
Gvalve
1/ 2
Pn
dP
G K d ρn 2
P ρ
o
Discharge Coefficient ( K d )
Values given by manufacturer, or in the “Red Book”
K d ,gas
If flow is not choked (sub-critical) use :
K d ,liquid
TWO-PHASE DENSITY
ρ ερG 1 ε ρL
Where ε is the volume fraction of gas:
x
ε
x S 1 x ρG / ρL
L
x xo xe xo
10
L = nozzle length (cm)
xo = initial quality entering nozzle
x e = local quality assuming equilibrium
If xo > 0.05, x = xe
Determine Quality, xe fn( P )
• The quality is determined as a function
of pressure by an energy balance on
the fluid along the flow path.
1/ 2 1/ 2
Pn
dP j n-1 Pj 1 - Pj
Gn ρn K d -2
P ρ
ρn K d -4
o j o ρj1 ρj
Required Information:
ρ vs P at constant s from Po to Pn in increments of
Pj to Pj+1 .
Can be generated from an EOS or from a
database (e.g. steam tables).
(If choked, Gn Gmax at Pn=Pc)
Experimental Data
TABLE I
VALVE SPECIFICATIONS
(Lenzing, et al, 1997, 1998)
Valve KdG KdL Orifice Dia. Orifice Area
(mm)
B&R DN25/40 0.86 0.66 20 0.4869
(Bopp &
Reuther Si63)
Data 5 bar
25000
Calc 8 bar
15000
10000
5000
0
0.0001 0.001 0.01 0.1 1
xo
Calc 5 bar
LESER DN25/40, Air/Water
Data 5 bar
25000 Calc 8 bar
Data 8 bar
20000 Calc 10 bar
K dG (kg/s m 2)
Data 10 bar
15000
10000
5000
0
0.0001 0.001 0.01 0.1 1
xo
Calc 5 bar
B&R DN25/40, Air/Water
Data 5 bar
Calc 8 bar
35000
Data 8 bar
30000 Calc 10 bar
Data 10 bar
25000
K dG (kg/s m2)
20000
15000
10000
5000
0
0.0001 0.001 0.01 0.1 1
xo
Crosby 1x2 E, Air/Water Calc 5 bar
Data 5 bar
25000
20000
K dG (kg/s m 2)
15000
10000
5000
0
0.0001 0.001 0.01 0.1 1
xo
HNDI – Homogeneous Non-Equilibrium
Direct Integration
For flashing flows, equilibrium is not reached
until flow path length reaches 10 cm or more.
For L<10 cm, quality (x = gas mass
fraction) is lower than it would be at
equilibrium (xe).
For L < 10 cm, quality is estimated from
x = xo + xe - xo L / 10
where xo is the initial (L = 0) quality (L in cm)
If xo > 0.05, x = xe
Steam-Water Flashing
(non-Equilibrium) Flow
5000
Data
4500 HDI
4000 HNDI L=40mm
KdG(kg/sm2)
3500
3000
2500
2000
1500
1000
500
0
0.001 0.01 0.1 1
xo
Leser Valve DN25/40, Steam/Water, 6.8 bar
7000
Data
6000 HDI
HNDI L=40mm
KdG(kg/sm2)
5000
4000
3000
2000
1000
0
0.001 0.01 0.1 1
xo
Leser Valve DN25/40, Steam/Water, 8 bar
8000
Data
7000 HDI
HNDI L=40mm
6000
KdG (kg/sm2)
5000
4000
3000
2000
1000
0
0.001 0.01 0.1 1
xo
Leser Valve DN25/40, Steam/Water, 10.6 bar
9000
Data
8000 HDI
7000 HNDI L=40mm
K dG (kg/sm 2)
6000
5000
4000
3000
2000
1000
0
0.001 0.01 0.1 1
xo
SUMMARY/MORAL
Two-Phase Flow is much more complex
than single –phase flow, because of the
wide variety of possible flow regimes,
phase distributions, thermo/mechanical
equilibrium/non-equilibrium, etc.
Darby, R., F.E. Self and V.H. Edwards, “Properly Size Pressure
Relief Valves for Two-Phase Gas/Liquid Flow”, Chemical
Engineering, 109, no. 6, pp 68-74, June, (2002)
Duckler, A.E, M. Wicks III and R.G. Cleveland, “Frictional Pressure Drop
in Two-Phase Flow: A Comparison of Existing Correlations for Pressure
Loss and Holdup, AIChE J., 10:38-43, 1964
Govier, G.W. and K. Aziz, “The Flow of Complex Mixtures in Pipes”, Van
Nostrand Reinhold Co., 1972