Section 2

Multiphase Flow, Flowing Well Performance

Multiphase Vertical Flow

When producing an oil or gas well, the flow of the fluids up the tubing will
be in most cases be 2 phase, liquid and gas. The pressure at any point in the
tubing string is influenced by many factors. Such as type of flow at that
point, the viscositys of the oil and gas, the PVT characteristics of the fluid
which are pressure and temperature dependent, the bottom hole and tubing
pressures. In the tubing there is a pressure drop and a temperature variation
which causes several of the fluid properties to vary along with the pressure
drop that allows the solution gas to come out of the oil at different rates in
various parts of the tubing, causing holdup behavior. The drag of the
flowing fluids (friction) also plays a part along with gravity in resisting the
flow.
Flow Regimes
As the oil travels up the tubing different types of flow can occur as the gas is
liberated from the liquid phase.
If the well is producing above the bubble point pressure the flow is single
phase but as it rises up the tubing the pressure will drop below this pressure
and the multiphase regimes will start to form.
Bubble Flow
The gas starts to come out of solution and form in the liquid medium
in an even dispersion.

Slug or Plug Flow

As the fluid moves upward the gas bubbles move faster than the
liquid, because of the buoyancy effect, and collect in large bubbles. These
bubbles grow to a size where extend across the diameter of the tubing
separating slugs of liquid containing the smaller bubbles

Churn Flow
With increasing gas velocity, the larger bubbles become unstable and
collapse, resulting in a highly turbulent flow pattern. Churn flow is
characterized by oscillatory, up and down motions of the liquid.

Annular Flow
As the pressure is lowered these large pockets of case break through
the slugs and form a continuos phase in the center of the tubing carrying
droplets of the liquid with a film of oil flowing along the walls of the tubing.

Mist Flow
Eventually the gas volume is so large that the film of liquid on the
wall of the tubing disappears and the only liquid moving is the droplets or
mist in the gas phase.

Not all of these regimes are necessarily present in a well but usually more
than one is present.
As can be seen the flow of oil and gas up a string of tubing has many
complex components that must be considered in designing a flowing well.
For a given bottom hole pressure there is a certain amount of reservoir fluids
that will enter the well bore, what is needed to be known is whether the
pressure drop in the tubing will allow flow to the surface. Several methods
have been developed to describe this flow so pressure curves of the tubing
can be drawn to use in the planning of the well and its equipment.
Poettmann & Carpenter Method
Based on the energy equation with the assumptions that the difference in the
kinetic energy at the top and bottom is negligible and that external work
done by the flowing fluid is also negligible. The energy loss Wf is caused by
the slippage and frictional effects.
W f = 4 fu 2

h
2 gD

(2-1)

where u is the average mixture velocity over h and D is the inside

diameter of the tubing, g is gravity constant and f is found empirically.
With the given assumptions and field units equation 2-1 can be reduced to

p _ K
144
= + _
h

where

(2-2)

p = pressure drop over h, psi

= average density of the fluid, lb/cu ft
fq 2 M 2
K=
(7.413x1010 D 5 )
_

q =
M=
D =
f =

(2-3)

liquid production rate, STB/day

Total mass of gas and liquid associated with 1 STB, lb
ID of tubing, ft
energy loss factor defined by equation 2-1

The way that this is used is to start at the bottom of the tubing with the
known pwf which can produce a certain production of reservoir fluids that is
known from the productivity index of the well. All the PVT and reservoir
parameters such as FVF, water-cut, gas solubility are known or can be
estimated. Divide the tubing into equal parts, H1, H2, H3, H4, each a length
of h and let the pressures, p1, p2, p3, p4, at the Hn points. At H1 the pressure
is pwf, so this pressure can be used for H1H2 for the first approximation for
the factors needed for equation 2-2, such as the density and Mass. If a better
approximation is needed use the average between of pwf and the first
approximation for the next calculations. This process is repeated for all the
intervals in the tubing, to the surface or till a negative pressure is calculated.
Doing this for several starting pwf a plot can be generated for any size tubing.

Figure 2-1
This graph can be used to find the bottom hole pressure and production rates
for any producing tubing head pressure. The maximum rate that the well
can flow is indicated by point E on the graph, this point must always be
lower than q since there is a positive pressure loss in the tubing.

Figure 2-2
Plotting the tubing head pressures for several sizes of tubing is used to find
the optimal tubing size for different flow rates. In this case at 200 psi THP
the 2 tubing is better but at 100 psi the 3 tubing produces at a higher
rate.
The only factor in equation 2-3 that is not easily found is the energy loss
factor f. In an effort to establish values for f as a function of the variables
involved in the flow of oil wells a reverse calculation was made. This
involved taking field data in the form of production, PVT data, and pressure
and temperature transverses. It was found that f correlated best as a function
of tubing ID and mass velocity of the fluid in the tubing. In terms of field
units
Dv = (1.4737 x10 5 )

Mq
D

From this the following plot is used to find f for different flow rates and
tubing sizes.

(2-4)

Figure 2-3
It is noted here that this plotted on log-log paper and that the scatter of the
points is greater that its appears on this graph. Also Baxendell extended this
work to in include flow in an annulus, and with Thomas plotted f over higher
flow rates.
The weakness of this approach is the inaccuracy of finding f. Work by
others have been to overcome this problem.

Using the Poettman & Carpenter Method

Calculating M, mass of the fluid in STB

M = wt of Oil + wt of Water + wt of Gas

in lbm/STB

wt of Oil = Oil cut x o x 350 lb/bbl

wt of Water = Water cut x w x 350 lb/bbl
wt of gas = GLR cf/bbl x g x .0764 lb/cf
w = Mass flow rate, = Mq, lbm / day
Note:

o =

141.5
131.5 + API o

w and g and GOR should be given &

the oil and water cut calculated from the WOR.

Calculating ave. Density, , lbm/cf
= M/volume
Volume = vol of oil + vol of water + vol of free gas in cf
Using the pressure of interval being calculated, lab or correlations of
PVT data needed. o, g, Rs
Volume bbls = Oil cut x o + Water cut + g{ GLR Oil cut x Rs}
Volume cf = Volume bbls x 5.614 cf/bbl
= M/(5.614[oil cut x o + water cut] + g{GLR Oil cut x Rs})

With values of M and the p/h can be calculated using equations 2-2
and 2-3 by finding f from figure 2-3. By dividing the pressure increment by
this pressure gradient gives the length of tubing over which the increment
occurs. This way a plot of the pressure can be obtained.

Example:
Given
qo. = 400 bpd
qw.= 600 bpd
GLR = 500 cf/bbl

Tubing 2
22o API oil
w = 1.07
g = .65
o = 141.5/131.5 + 22 = .922

M = .4*.922*350 + .6*1.07*350 + 500*.65*.0764 = 378.6 lb/bbl

w = M*q = 378.6*1000 = 378610 lb/day
Dv = 1.4737E-5 x w/D = 1.4737E-5 * 378610/1.995/12 = 33.56
Form fig. 2-3 f = .008
K = f w2 /(7.413E 10 D5) = .008*378610^2 / (7.413E+10*.166^5) = 122.7
For pressure point 1, 500 psi
o = 1.043
Rs = 59 scf/bbl

z = .93

= M/5.614[oil cut x o + water cut] + g{GLR Oil cut x Rs}

378/(5.614[.4*1.043 + .6] + .0298{500 - .4*59}]) = 19 lb/cf
For pressure point 2, 1000 psi
o = 1.083
Rs = 120 scf/bbl z = .9
= 378.6/(5.614[.4*1.083 + .6] + .0153{500 - .4*120}) = 29.8 lb/cf

ave = 1 + 2 / 2 = 24.4 lb/cf

Use this average density to find p/h in equation 2-2
p/h = [ + K/]/144 = [24.4 + 122.7/24.4]/144 = .204 psi/ft

Now solve for /h for the p

h = 1000 500 / .204 = 2447 ft.
Plot this point, say THP is 500 psi then it 1000 psi at a depth of 2447.
Then repeat the process for the next interval.

A large interval was used here in the example, the interval used when doing
a pressure profile of a well the pressure intervals will be much smaller, like
about 25 psi.

Modified Hagedron and Brown Method (mH-B)

This is an empirical two-phase flow correlation, the core of which is
correlation for liquid hold-up. Griffith correlation is used for flow in the
bubble flow region.
The mechanical energy balance equation used for the mH-B correlation
144

_
(um2 / 2 g c )
dp _
fw 2
= +
+

_
dh
h
(7.413E10D 5 )

(2-5)

in oilfield units.Bubble flow regime exists if g < LB where

u2
LB = 1.071 .2218 m
D

(2-6)

if LB <.13 set LB to .13.

The input fraction of gas g is g =

u sg
um

To find the liquid hold-up that is needed to calculate the pressure gradient
from correlations require the following dimensionless numbers
Liquid velocity

N vl = 1.938u sl 4

(2-7)

Gas velocity

N vg = 1.938u sg 4

(2-8)

Pipe diameter

N D = 120.872 D

(2-9)

N L = .15726 l 4

Liquid viscosity

l 3

(2-10)

Where superficial velocities us in ft/sec, density in lbm/ft3, surface tension

in dynes/cm, diameter D in feet.

Using mH-B method

First find the flow regime.
Calculate mass flow rate w and density as before.
Calculate the dimensionless numbers.
Calculate the velocities ul & ug
ul =

ql 5.614
At 86400

ug =

ql GLR g
At 86400

um = ul + u g

Using NL find CNL from fig. 2-3

N vl p 0.1 (CN L )
N vg0.575 pa0.1 N D

Then find the group

find yl/ from fig 2-4

p here is the absolute pressure at the location of interest.

Compute

N vg N L0.38
N D2.14

from fig 2-5 we get

Get the liquid hold-up yl

yl =(yl / )/

Note: For low viscosity fluid will generally be 1.

For f a 2 phase Reynolds number is be used

N Re =

2.2 E 2 w
Dlyl g(1 yl )

f from Moody diagram.

Get pressure gradient from, assuming no kinetic energy
dp _
fw 2
144
= +
_
dh
(7.413E10D 5 )

The Griffith correlation

For bubble flow Griffith uses a different hold-up correlation, bases the
frictional pressure gradient on the in-situ average liquid velocity.
144

fwl2
dp _
= +
(7.413E10)D 5 l yl2
dh

(2-11)

The liquid hold-up is

2

um
1 um
4 sg
yl = 1 1 +
1 +
2
us
us
us

(2-12)

where slip velocity us is .8 ft/sec.

Reynolds number is based on in-situ liquid velocity
_

N Re =

D ul l

or N Re =

2.2 E 2 wl
D l

(2-13)

figure 2-3

figure 2-4

figure 2-5

Equations to replace figure 2-3

(CN L ) = 10Y

Y = 2.69851 + 0.15841X 1 0.55100 X 12 + 0.54785 X 13 0.12195 X 14

X 1 = log[(N L ) + 3]

to replace figure 2-4

( y L / ) = 0.10307 + 0.61777[log( X 2 ) + 6] 0.63295[log( X 2 ) + 6]2 +
+ 0.29598 [log( X 2 ) + 6] 0.0401 [log( X 2 ) + 6]
3

X2 =

N vL p 0.1 (CN L )
0.575 0.1
N vG
pa N D

for figure 2-5

= 0.91163 4.82176X 3 + 1232.25 X 32 22253.6 X 33 + 116174.3 X 34
N vG N L0.38
X3 =
N D2.14

if X3>.01
for X3<.01 then = 1
and for f

1.1098 7.149 0.8981
1
5.0452

log
= 4 log

N Re
ff

2.8257 N Re
3.7065

Find the pressure gradient at 500 psi

qo. = 400 bpd
qw.= 600 bpd
GLR = 500 cf/bbl
o = 30 dynes/cm
o = 3 cp

Tubing 2
22o API oil
w = 1.07
g = .65
w = 70 dynes/cm
g = .015 cp
o = 141.5/131.5 + 22 = .922

M = .4*.922*350 + .6*1.07*350 + 500*.65*.0764 = 378.6 lb/bbl

w = M*q = 378.6*1000 = 378610 lb/day
Find the flow regime
ql GLR g

ul =

ql 5.614
At 86400

ul =

1000 5.614
= 2.98 ft / sec
.022 86400

ug =

um = ul + u g

At 86400

ug =

1000 500 .0298

.022 86400

= 7.84 ft / sec

u m = u l + u g = 2.98 + 7.84 = 10.82 ft / sec

u m2
LB = 1.071 .2218
D

LB<.13 so LB =.13

10.82 2
= 1.071 .2218

.166

g =

= 155.3

u sg
um

= 7.84 / 10.82 = .725

g is larger than LB so the flow regime is not bubble.

l = Ml/5.614[oil cut x o + water cut]
353.8/5.614[.4*1.043 + .6] = 61 lb/cf
g = Mg/ g{GLR Oil cut x Rs}
24.8/ .0298{500 - .4*59}=1.75 lb/cf

N vl = 1.938u sl 4

l
61
= 1.938 2.984
= 6.89
54

N vg = 1.938u sg 4

l
61
= 1.938 7.844
= 15.66
54

N D = 120.872 D

N L = .15726 l 4

l
61
= 120.872 .166
= 21.4
54

= .15726 2.984

1
= .0084
61 54 3

From chart 2-3 CNL = .0022

N vl p 0.1 (CN L )
6.89 500 .1.0022
=
= 2.1E 4
N vg0.575 p a0.1 N D 15.66 .57514.7 .1 21.4

From chart 2-4

N vg N L0.38
N D2.14

15.66 .0084 .38

= .0036
21.4 2.14

From chart 2-5

N Re =

yl/ = .47

=1 so yl = .47

2.2 E 2w
2.2 E 2 378610
=
= 2777288
yl
(1 yl )
.166 3.47 .015 (1.47 )
D l g

From the graph f = .005

144

dp _
= +
dh

fw 2

(7.413E10 D )
5

= 19 +

.005 378610 2
= 3.63 / 144 = .028 psi / ft
7.412 E10 .166 519

Homework
Use the data from the previous homework and solve the pressure gradient at
1000 psi.
Additional data
o = 4 cp
g = .02 cp
o = 30 dynes/cm
w = 70 dynes/cm

Ros Method
Uses a pressure balance rather than a energy balance equation. So that in a
single phase flow the pressure gradient is the sum of the static, friction and
acceleration gradients,
dp
= g + 4 f
dh

1 v 2
2
D

(2-14)

here with the acceleration gradient neglected.

For 2 phase flow the static term should be modified to allow for gas slippage
though the liquid. Ros introduced the a static gradient
l l g + g g g

(2-15)

Where is the liquid and gas holdup. Holdup is defined as the volume of
that phase actually present in a certain length of pipe.
By substituting equation 2-6 into 2-5
g
1 dp
= l + (1 + l )
+ friction term
l
l g dh

(2-16)

The density of gas is small compared to the oil so 2-7 can be rewritten;
1 dp
= l + friction term
l g dh

(2-17)

Ros then conducted experiments to determine the influence of varies factors,

resulting in an empirical correlations for liquid holdup and wall friction
which are very complex. A computer program based on these correlations
calculate pressure gradients that compare favorably with Baxendell &
Thomas field data.

Gilbert Method

Empirical based method, based on measured values of tubing flow pressure

losses, a families of curves were derived that can be used for extrapolation
and interpolation purposes.
The following measurements were taken on a large number of wells;
Tubing depth, ft
Flowing BHP, tubing in-take pressure, psi
THP, psi
Flow rate, gross liquid, bbl/day
GLR, mcf/bbl
Tubing size
Assuming that the BHP depends only on the other five listed variables,
group the wells with the same tubing size, production rate, and GLR to get a
correlation. Plot the BHP as a function of depth for these groups and a plot
like figure 2-4 is obtained showing the pressure curves for several THP.

figure 2-6

Take a point on the curve of a THP A at a selected depth and place a valve
there that can be opened in such a way that the pressure and production rate
will be unchanged when the valve at the tubing head is closed. It can be
seen that pressure curve below this point will also be unchanged and that this
point can now be viewed as the tubing head. So actually all the curves are
just extensions of one curve. This curve is given in figure 2-6.

figure 2-7
To use this curve to find the BHP at a given depth, the equivalent depth of
the THP is found on the curve and added to the total length of the tubing.
For example if the THP is 200 psi, and the BHP at 10,000 is wanted, the
depth on the curve of 200 psi is noted, say is 1200, this is added to the
10,000, making a depth of 11,200 which is used on the curve to find the
BHP.
It should be noted that at low pressures it found that the correct shape of the
curve is actually concave downward or the reverse of the higher pressure
portion of the curve. This is seen if Poettmann and Carpenter method is
used to calculate the curves. Baxendell and Carpenter ran very carefully
controlled well tests and confirmed this shape at the lower pressures.

The Gilbert Curves use GLR as a parameter and there is one family of
curves for each tubing size and flow rate.

Two approaches to solving flowing well problems.

1) Find the IPR curve of the well, then find the pwf for different flow rates
for a given THP. The intersection of the plot of these two lines will be the
expected flow rate for the given THP.

IPR & 200 THP

IPR
thp 200

1500
1400
1300

Pwf psi

1200
1100
1000
900
800
700
600
0

100

200

300

400

qo bpd

500

600

700

2) After calculating the IPR curve, calculate the pressure loss in the tubing at
various flow rates. This will give a plot like figure 2-1, it can be seen that
for any flow rate a THP and BHP (pwf) can be found. Also the flow rate for
and THP can be found.