Sie sind auf Seite 1von 33

40

2.2 Tubing performance analysis


Performance of tubing is directly related to the summation of pressure drop in the
completion system which includes:
production tubing,
sub-surface choke,
surface choke,
flow line,
separator and
gathering line to sale point.

In a single phase flow, the calculation of pressure loss is relatively simple and the
pressure values can be determined accurately by using well established methods. For
multiphase flow same procedure can be applied, however, the parameters for pressure
terms can not be accurately determined at an acceptable level. Problem of calculating
pressure loss in multi phase flow is that the phase behaviour and flow pattern are
temperature and pressure dependant which can vary from bottomhole to the surface. The
flow from reservoir to the well is single phase as long as the reservoir pressure remains
above the bubble point pressure. Due to the change in pressure and temperature inside
tubing, the pressure might fall below the bubble point. This will result in liberation of gas
from liquid (oil). As pressure continues to drop, the gas phase expands and additional gas
comes out of the solution. This could change fluid stream from single phase liquid at the
bottomhole to mostly gas phase at the near surface. Typical flow regimes in a tubing are
illustrated in Fig. 2.7.

DISTRIBUTED INTERMITTENT SEGREGATED


MIST
STRATIFIED
ANNULAR
BUBBLE

SLUG

PLUG

WAVY

Fig. 2.7: Flow regimes, after Breggs and Bill 1978

School of Petroleum Engineering, UNSW


41

Empirical and semi-empirical analysis:

The process of estimating pressure loss when multiphase flow exists through tubing is
complex. Because of this complexity, empirical and semi-empirical analysis techniques
have been used to develop relationships among the producing conditions listed in Fig.2.7.
There are a number of correlations incorporated in commercial software (e.g.
VIRTUWELLT M) or published as gradient curves. Since these correlations give
somewhat different results, the engineer should establish a match with field test data and
choose the most appropriate correlation.

Three of the most commonly used correlations are:


Hagedorn and Brown,
Griffith and
Beggs and Brill.

Hagedorn & Brown / Griffith:

An effort was made by Hagedorn and Brown to determine a correlation which would
include all practical ranges of flow rate, a wide range of gas-liquid ratios, all ordinary
used tubing sizes and the effects of fluid properties. The heart of the Hagedorn and
Brown correlation is a correlation for liquid holdup. This correlation is selected based on
the flow regime as follows. Bubble flow exists if g < LB where:
vm2
LB 1.071 0.2218( ) (2.18)
d
v m is the mixture velocity, g is the input gas fraction and is:
qg
g (2.19)
q g ql
If the value of LB is less than 0.13, it is set to 0.13. If the flow regime is found to be
bubble flow the Griffith correlation is used otherwise the Hagedorn and Brown
correlation is used.
Flow Regimes other than bubble flow:
The Hagedorn and Brown correlation is an empirical two phase flow correlation:
dp g 2 f f vm2
m cos (2.20)
dh gc gc d
Where,
dp
= Pressure gradient (lb/ft3 ),
dh
f = friction factor,
= angle of well segment from vertical,
d = diameter (ft),
m = mixture density (lb/ft3 ),
g= acceleration due to gravity, 32.2 ft/sec2 ,
gc= conversion factor in Newtons second law of motion, 32.2 lb m ft/(lbfsec2 ),
L= no-slip liquid hold-up,

School of Petroleum Engineering, UNSW


42

n = no-slip mixture density, lb/ft3 and


n2
f .
m
The mixture velocity is the sum of the superficial velocities of each phase (v sg and v sl)
which can be calculated using:

v m v sg v sl (2.21)
qg
v sg (2.22)
A
and
ql
v sl (2.23)
A

Where,
qg and ql are the gas and liquid flow rates respectively.
The no-slip liquid hold-up and density are calculated as follows:
v
L sl (2.24)
vm
n L L g 1 L (2.25)
where L and g are liquid and gas densities respectively.

The liquid holdup is obtained from a correlation and the friction factor is based on a
mixture Reynolds number. Using the following dimensionless numbers we can determine
the liquid holdup from a series of charts and are defined as:

Liquid velocity number:


l
N vl 1.938v sl 4 (2.26)

Gas velocity number:
l
N vg 1.938vsg 4 (2.27)

Pipe diameter number:
l
N D 120.872d (2.28)

Liquid viscosity number:
1
N L 0.15726 l 4 (2.29)
l 3
Where,
Superficial velocities are in ft/sec
Density in lbm/ft3 ,
Surface tension, ( ) in dynes/cm viscosity in cp and diameter in ft.
To obtain holdup first calculate NL and read the value of CNL from Fig. 2.8.

School of Petroleum Engineering, UNSW


43

Fig. 2.8 Hagedorn and Brown correlation for CNL (from Hagedorn and Brown, 1965)

Then the following group is calculated


N vl p 0.1 (CN L )
0.575 0.1
(2.30)
N vg pa N D
P is the absolute pressure at the location where the pressure gradient is wanted and pa is
the atmospheric pressure
H
A value for L , where HL is the holdup factor, can be obtained using this group and

Fig. 2.9.

yl
Fig. 2.9 Hagedorn and Brown correlation for (from Hagedorn and Brown, 1965)

School of Petroleum Engineering, UNSW


44

Finally calculate the following group and use it and Fig. 2.10 to get
N vg N L0.380
(2.31)
N D2.14

Fig. 2.10 Hagedorn and Brown correlation for (from Hagedorn and Brown, 1965)

The liquid holdup is defined as:


H
H L ( L ) (2.32)

The mixture density is calculated below as :
m L H L g 1 H L (2.33)

The frictional pressure gradient is based on the Fanning friction factor. To obtain this
value Reynolds number must be calculated:
1488 n v m d
N Re m (2.34)
m
Where, mixture viscosity, cp, is defined as:
m LH L g1 H L (2.35)
The friction factor is obtained using the Moody diagram or from the following equation:

2


1
f (2.36)
21.25

1.14 2 log d N 0.9
Re m
Where,
= pipe roughness, ft

School of Petroleum Engineering, UNSW


45

Bubble flow, the Griffith correlation:


The Griffith correlation uses Eq 2.37 to calculate pressure gradient.
dp fm l2
144 (2.37)
dz 7.413 1010 d 5 l y l2
Where m l is the mass flow rate of the liquid only and v l is the in situ average liquid
velocity defined as:
v sl q
vl l (2.38)
yl Ay l

1 vm v v sg
yl 1 1 (1 m ) 2 4 (2.39)
2 vs vs vs

Where v s = 0.8 ft/sec (v s is the slip velocity). Reynolds number can be calculated using:
2.2 10 2 m
N Re (2.40)
d l

Beggs and Brill


The Beggs and Brill correlation was developed from experimental data obtained in a
small scale test facility. It differs significantly from Hagedorn and Brown. Beggs and
Brill correlation is applicable to any pipe inclination and flow direction. The overall
pressure gradient can be calculated using:

dp dp
) PE ( ) F
(
dp
( ) dh dh (2.41)
dh 1 Ek

The kinetic energy contribution to the equation is accounted for by the Ek parameter:

v m v sg m
Ek (2.42)
gc p
The potential energy pressure gradient is:
dp g
( ) PE s sin (2.43)
dh gc
and the frictional pressure gradient can be calculated using:
dp 2 ftp n vm2
( )F (2.44)
dh gc d

Where,
n l l g g , (2.45)
f tp = Two phase friction factor,
v sg = Superficial gas velocity (ft/sec),
= The angle between the horizontal and direction of flow ,
l = Input liquid fraction,

School of Petroleum Engineering, UNSW


46

g = Input gas fraction,


g = gravitational acceleration (32.2 ft/s2 ) and
s l H L g 1 H L (2.46)
The two phase friction factor, f tp is determined using:
f tp f n e S (2.47)
Where,
S ln 2.2 x 1.2 for 1<x<1.2 (2.48)

Otherwise,
ln x
S
0.0523 3.182 ln( x) 0.8725ln( x) 0.01853ln( x)
2 4
(2.49)

and
l
x (2.50)
H L2

The no slip friction factor, f n is based on a smooth pipe ( 0 ) and the Reynolds number:
d

n vm d
N Re m (2.51)
n
Where,
n l l g g (2.52)
The correlation is based on the following parameters:
v m2
N FR (2.53)
gd
v sl
l (2.54)
vm
L1 3160l .302 (2.55)
L2 0.0009252l2.4684 (2.56)
L3 0.10l1.4516 (2.57)
L4 0.5l6.738 (2.58)

The flow regimes are given by the following:


Segregated flow exists when:
l < 0.01 and N FR < L1 or l 0.01 and N FR L2 (2.59)
Transition flow exists when:
l 0.01 and L2 < N FR L3 (2.60)
Intermittent flow exists when:
0.01 l < 0.4 and L3 < N FR L1 or l 0.4 and L3 < N FR L4 (2.61)
Distributed flow exists when:
l < 0.4 and N FR L1 or l 0.4 and N FR > L4 (2.62)

School of Petroleum Engineering, UNSW


47

For segregated, intermittent and distributed flow the following equations are used to
calculate liquid holdup and hence average density:
H L ( H L ( 0) ) (2.63)
abl
H L ( 0) c
( H L ( 0) must be L ) (2.64)
N FR

1 C sin(1.8 ) 0.333 sin 3 (1.8 ) (2.65)


C 1 l ln del N vlf N FR
g
(2.66)

Where a, b, c, d, e, f and g are dependent on the flow regime. Values for these constants
are specified in Table 2.1.

Table 2.1 Beggs and Brill Holdup Constants

Beggs & Brill Holdup Constants

Flow Regime a b c
Segregated 0.98 0.4846 0.0868
Intermittent 0.845 0.5351 0.0173
Distributed 1.065 0.5824 0.0609
Flow Regime d e f g
Segregated Uphill 0.011 -3.5868 3.519 -1.614
Intermittent Uphill 2.96 0.305 -0.4473 0.0978
Distributed Uphill No Correlation C=0, =1
All regimes dowhill 4.7 -0.3692 0.1244 -0.5056

For transition flow the liquid holdup is calculated using both the segregated and the
intermittent equations and interpolated using the following:

H L AH L (segregated ) BH L (intermitte nt ) (2.67)

Where,

L3 N FR
A (2.68)
L3 L2

and

B 1 A (2.69)

School of Petroleum Engineering, UNSW


48

These correlations are used to calculate pressure gradient which can be applied to a well
at random locations. However our objective is to calculate the overall pressure drop p
over a considerable distance. Over large distances the pressure gradient in two phase flow
will vary significantly as the downhole flow properties change with temperature and
pressure. For example in a single phase flow oil well, if pressure drops below the bubble
point gas comes out of solution. This will cause gas-liquid bubble flow and as the
pressure continues to drop other flow regimes may occur farther up the tubing. Since both
temperature and pressure are varying over the length of the tubing, the calculation for
total pressure drop will generally be iterative. An algorithm to calculate the pressure loss
using the empirical methods which are explained above is illustrated in Fig. 2.11

School of Petroleum Engineering, UNSW


49

START

Divide the pipe into segments Calculate Pave and Tave

For n = 1, # of segments Calculate PVT properties = f(Pave,Tave)

Initialize: Define Pinlet Calculate flow rates and densities


1

Calculate in-situ superficial


Assume dP for segment
velocities (Eqn. 2.21 - Eqn. 2.23

Beggs Hagedorn &


Calculate parameters used & Brill Brown/Griffith
Select
to determine flow pattern Calculate Eqn. 2.18,2.19
correlation
(Eqns. 2.53 - 2.58)

YES NO
Is g < LB ?

Determine flow pattern


(Eqns. 2.59 - 2.62)
Use Griffith Use Hagedorn & Brown
correlation correlation

Segregated, Transition Flow Calculate in-situ Calculate dimensionless


Intermittent & average liquid velocity numbers (Eqns. 2.26 -
Distributed Flow (Eqn. 2.38) 2.29)
Calculate HL for
segregated and
intermittent flows (Eqn. Determine CNL from
Calculate HL from Eqn. Calculate NRe
2.63 - 2.66; Table 2.1) Figure 2.11
2.63 - 2.66. Values of (Eqn. 2.40
a,b,c,d,e,f,g are found
in Table 2.1 )
Calculate Eqn. 2.30 &
determine HL/from
Interpolate using
Eqn. 2.67 - 2.69
Calculate the actual and Fig. 2.12
Calculate f
no-slip densities (Eqn.
(Eqn. 2.36
2.45 & Eqn. 2.46) Calculate Eqn. 2.31 &
determine from Fig.
2.13
Calculate x
(Eqn. 2.50) Calculate dP Calculate HL from
NO (Eqn. 2.37) Eqn. 2.32
YES
Calculate S from Calculate S from
Is 1<x<1.2
Eqn. 2.49 Eqn. 2.48
Calculate mixture
density (Eqn. 2.33)

Calculate ftp from


Eqn. 2.47
Determine NRem
(Eqn. 2.34)

Calculate dP from Calculate dP (Eqn. Calculate friction


Eqn. 2.41 2.20) factor (Eqn. 2.36

NO
|dPassumed-dPcalculated|<TOL? dPassumed=dPcalculated 1

YES
L<Total Pinlet=Pinlet+dP
2
length? L=Li+dL

STOP

Fig. 2.11 Algorithm for pressure traverse calculation using either Hagedorn
& Brown/Griffith or Beggs & Brill correlation.

School of Petroleum Engineering, UNSW


50

EFFECTS OF SOME IMPORTANT PARAMETERS ON BOTTOMHOLE


FLOWING PRESSURE

Parameters affecting pressure loss inside tubing include:

tubing diameter (D),


flow rate (q),
gas-liquid ratio (GLR),
water cut,
fluid density ,
viscosity ,
pressure and
temperature

Effect of GLR:
Unlike single phase reservoirs, the GLR (gas liquid ratio) will vary with time as the
pressure in the reservoir changes. As GLR increases, the density of produced fluid
decreases, which will result in decrease in BHP (bottomhole pressure).
Usage of GLR curves is illustrated in Fig. 2.12. The important thing to remember is enter
the curve at a point defined by the rate, GLR and flowing tubing pressure, or BHP (THP
equivalent to 1,000 ft in Fig. 2.12), and then move along the appropriate GLR line by an
increment equivalent to the depth (i.e., from 1,000 to 8,000 ft for a 7,000-ft deep well).
Do not just read the BHP conditions at a given depth - this merely corresponds to a value
of 0 THP. The other important considerations are that you use the correct water cut and
adjust the GOR to a GLR:

GLR = (1 WC) x GOR (2.17)

For deviated wells, it may be necessary to use a computer or to interpolate between true
vertical depth and measured depth by deducting the additional head effects using an
average effective density.

School of Petroleum Engineering, UNSW


51

Fig. 2.12: Vertical flowing pressure gradient, (from Brown, 1982)

Effect of Liquid Density:


As liquid density decreases, required flowing bottomhole pressure decreases. Fig. 2.13
compares the effect of API gravity crude to fresh and salt water.

School of Petroleum Engineering, UNSW


52

Fig. 2.13: Effect of API gravity on required flowing BHP, (from Brown, 1982).

Effect of Liquid Viscosity:


As liquid viscosity increases, higher flowing bottomhole pressures are required.

Effect of Liquid Surface Tension:


The required flowing bottomhole pressure increases with increased surface tension.

Effect of Kinetic Energy:


The kinetic energy effect on flowing pressure can become important for small diameter
tubing with high gas/liquid ratios and low pressure levels.

Effect of Subsurface and Surface Choke Size:


These two chokes contribute to substantial amount of pressure loss within the producing
equipment. Subsurface chokes are installed at specific depths depending on their
functioning .The pressure loss at subsurface choke is comparatively greater when they are
installed at a shallow depth than at deeper locations.
To study the effect of surface chokes on pressure drops, a plot of pressure drop vs
upstream pressures is plotted with different sizes of chokes. It is observed that the
pressure drop decreases with increase in the upstream pressures for a constant flow rate.
This is illustrated in Fig. 2.14.

Effect of Flow line:


At a specific length it is observed that as the size of the flow line increases the pressure
drop decreases.

School of Petroleum Engineering, UNSW


53

Effect of Separator pressure:


Usually this pressure is maintained constant level at the separator. As the separator back
pressure varies at given conditions, it will alter the flow rate substantially.
For more detail study in regards to effects of individual factors mentioned above please
refer Production operations -1, by Allen and Roberts (1989) and Production operations
course -1 by L.E. Buzarde et.al.(1972)

Fig. 2.14: Effect of surface choke size and upstream pressure, (from Brown, 1982).

School of Petroleum Engineering, UNSW


54

VERTICAL LIFT PERFORMANCE ANALYSIS


In process of well design a plot of flowing bottomhole pressure (pwf) versus rate (q) for
various tubing sizes and gas liquid ratios would be most useful in presenting the
performance of the tubing.
For a selected tubing size, there is a minimum flow rate which is required to attain
production of hydrocarbons from the well. This is the rollover point in the tubing
performance curves, which is not easily recognized without a computer simulation. In
practice a fluid velocity of at least 5 ft/s is required for production. Below this rate the
well will be unstable. This phenomenon is referred to as liquid holdup and is due to
slippage of the gas phase through the liquid. The larger the tubing diameter, the higher
will be the liquid holdup rate.

Matching Completion and Reservoir Performance:


The objective of this analysis is to calculate the maximum productivity of the system.
The output of the analysis is the complete drilling design. We determine a size of tubing
which can deliver maximum performance at a specific GOR. And thus allowing us to
determine the maximum size of the casing needed to complete the well. This analysis is
performed at the development drilling planning phase so that an adequate casing size is
planned. In obtaining this plot, the well head pressure and about 5 flow rates that are
likely to be within the extent of productivity are assumed.

The steps involved in determining the tubing size are given below:
I. Plot the IPR of the reservoir.
II. Plot the VLP of different tubing sizes.
III. The final step is to combine them and identify the intersection points that indicate
the maximum productivity of the system.
IV. While this requirement is obvious for flowing wells, gas-lift operations, and
injection wells, it is often forgotten when other artificial lift systems are used.
V. The production target rate and the expected water cut and GLR behavior will also
be constraining factors that will be evaluate.
VI. Combination of VLP(Vertical lift performance) and IPR

Example 2.5: Calculation of VLP for single phase fluid and combination with IPR
For a 8000 ft oil well (oil gravity, = 0.88) with a tubing I.D of 2 in and the following
properties what would be the expected production rate and corresponding bottomhole
pressure if the wellhead pressure is 0 psi? Assume the bubble point pressure is zero. The
reservoir operates under steady state conditions. Ignore any kinetic energy losses.

k = 8.2 md,
h = 53 ft,
pR = 5651 psi,
= 1.7 cp,
B = 1.1 rb/STB,
rw = 0.328 ft,
s = 0,

School of Petroleum Engineering, UNSW


55

re = 2980 ft,
= 55 lb/ft3 and
= 0.0006.
For steady state flow:
r
141.2 o Bo q ln e S

rw
p wf pR
kh
The IPR curve may be represented as:
p wf 5651 5.54q
To calculate the potential energy pressure loss we have:
PPE 0.433z
PPE 0.433(0.88)(8000) 3048 psi
Since the well is considered to be single phase and the fluid is considered to be
incompressible the potential energy drop will be the same regardless of flow rate.
Pressure losses due to friction can be calculated using:
2 f f v 2 z
PF
gcd
Where,
4q
v
d 2
and
1 5.0452 1.1098 7.149 0.8981
4 log( log( ( ) ))
ff 3.7065 N Re 2.8257 N Re

To construct a VLP curve we need to find p wf at different flow rates:


For q = 100 STB/day
1.48(100)(55)
N Re 2400
(2)(1.7)
f f 0.0117
2(0.0117)(55)( 0.3) 2 z (8000) lb f
PF 173 2 1.2 psi
2 ft
(32.17)( )
12
The total pressure drop will be the sum of the potential energy drop and the frictional
pressure loss. Since ptp = 0 this is also equal to flowing bottomhole pressure, pwf
p wf 3048 1.2 3049 psi
The following table provides the values for pwf at different flow rates using similar
calculations:

qo (STB/day) pwf (psi)


100 3049
300 3056
500 3068
700 3083

School of Petroleum Engineering, UNSW


56

Tutorial: Pressure Loss inside the tubing based on Hagedorn and Browns approach

Calculate the pressure loss inside the tubing based on Hagedorn and Browns approach
based on the following data. Ignore Kinetic pressure loss.

Production Rate, Qo 10,000 STB/day


Pressure @ a given point 1700 psi
Superficial Liquid Velocity, Vsl 3.97 ft/sec
Superficial Gas Velocity, Vsg 3.86 ft/sec
Oil Viscosity, o 0.97 cp
Oil Density, o 47.61 lb/ft3
Gas Viscosity, g 0.016 cp
Gas Density, g 5.88 lb/ft3
Gas/Oil Interfacial Tension, 8.41 dynes/cm
o(C)
Diameter of the pipe 6 inch
Surface Roughness, 0.0006 ft
L 10000 ft

Solution:

Mixture Velocity = V m = V sl+V sg


Vm = 7.83 ft/sec

Inside Area of pipe = Ai =


A= 0.196 ft2

Liquid Flowrate = ql = V sl x Ai
ql = 0.778 ft3 /sec

Gas Flowrate = qg = V sg x Ai
qg = 0.757 ft3 /sec

To find out Flow Regime:

LB=1.071-0.2218(V m2/d), d should be in ft


LB = -26.126

No Slip Gas Holdup = g = qg/(qg+ql)


g = 0.493

Note: if the value of L B is less than 0.13, it is set to be 0.13!


Therefore, LB = 0.13

School of Petroleum Engineering, UNSW


57

Type of Flow:
If g<LB then Bubble Flow (use Griffith Correlation).
If g>=LB then use Hagedorn & Brown Correlation OR Beggs & Brills Correlation.

Since the flow is NOT a bubble flow we use Hagedorn & Brown Correlation (Also this is the
requirement of this question).

HAGEDORN & BROWN CORRELATION

No-Slip Liquid Holdup = L = 1 - g


L = 0.507
No-Slip Mixture Density =
= 27.04 lb/ft3

Calculation of Dimensionless numbers:

Liquid Velocity Number = N vl = 1.938 VSL (L/ )


Nvl = 11.87

Gas Velocity Number = N vg = 1.938 Vsg (L/ )


Nvg = 11.54

Pipe Diameter Number = N D = 120.872 d (L/ ) , d is in ft


ND = 143.8

3
Liquid Viscosity Number = N L = 0.15726 L (1/(L )
NL = 0.0118

From Chart, Value of CN L corresponding to N L

School of Petroleum Engineering, UNSW


58

CNL

CNL = 0.0024

Let

a = 7.81x10-05

From Chart, Value of HL/ corresponding to a is:

HL/ = 0.3

Let
b = 5.1 x 10-05

School of Petroleum Engineering, UNSW


59

From Chart, Value of corresponding to b is:

=1
Liquid Holdup = HL = (HL/) *
HL = 0.3 x 1
HL = 0.3
Note: in Modified Hagedorn and Brown Method, if Liquid Holdup (HL) is less than no-
slip liquid hold-up (L) then Liquid Holdup will be set equal to no-slip liquid hold-up (HL
= L).
Mixture Density = m = LHL+g(1-HL)
m = 18.4 lb/ft3

Mixture Viscosity = m = LHL g(1-HL)


m = 0.0548 cp

Reynolds number = , d is in ft

N Rem = 2.87 x 1006


2


1
Friction factor = f , d is in ft
21.25

1.14 2 log d N 0.9
Re m
f = 0.02468

School of Petroleum Engineering, UNSW


60

f = n 2 /m
f = 39.74 lb/ft3

Pressure Loss inside Tubing = dp/dh


dp/dh = Pressure loss due to Potential Energy + Pressure loss due to Friction

Press loss due to Potential Energy=P PE


PPE = ( )m cos / 144

PPE = ( ) 18.4 x cos(0) / 144

PPE = 0.1278 psi/ft

Press loss due to Friction=Ploss friction

Ploss friction = /144

Ploss friction = 0.0519 psi/ft

Pressure Loss inside Tubing = dp/dh = PPE + Ploss friction


dp/dh = 0.18 psi/ft
Therefore Pressure Loss inside Tubing = 0.18 psi/ft x 10000 ft

Pressure Loss inside Tubing = 1800 psi

School of Petroleum Engineering, UNSW


61

2.3 Well Performance Analysis using Virtuwell TM

The calculation of pressure drops through different production scenarios can be very time
consuming. These calculations can be particularly tedious in cases in which multiphase
flow is expected. For this reason, many companies in the oil industry rely on computer
software to model and predict the pressure drops in tubulars for a given flow mixture.
The software can also be used to predict the productivity index and tubing performance.

One such program used in industry is VirtuwellT M. This software incorporates empirically
derived mathematical models to predict the flow behavior and pressure drop through
tubulars for single and multiphase flow. It can be used for vertical or deviated wells. The
correlations used will vary for the composition of the fluid. For this reason, it uses one set
of correlations for single-phase liquid or gas flow, and another set for multiphase flow.
The correlations used in VirtuwellT M to model single-phase and multiphase flows are now
discussed in the following section.

Virtuwell TM Single-Phase Flow Correlations:


Generally it is easier to calculate pressure drops for single-phase flow than it is for
multiphase flow. There are three single-phase correlations that are available and they are:

Fanning:
This correlation is divided into two sub categories Fanning Liquid and Fanning Gas. The
Fanning Gas correlation is also known as the Multi-Step Cullender and Smith when
applied for vertical well bores.

Panhandle:
This correlation was developed originally for single-phase flow of gas through horizontal
pipes. In other words, the hydrostatic pressure difference is not taken into account. We
have applied the standard hydrostatic head equation to the vertical elevation of the pipe to
account for the vertical component of pressure drop. Thus our implementation of the
Panhandle correlation includes both horizontal and vertical flow components, and this
equation can be used for horizontal, uphill and downhill flow.

Modified Panhandle:
This is a variation of the Panhandle correlation that was found to be better suited to some
transportation systems. Thus, it also originally did not account for vertical flow.
VirtuwellT M applies the standard hydrostatic head equation to account for the vertical
component of pressure drop. Hence, the implementation of the Modified Panhandle
correlation includes both horizontal and vertical flow components, and this equation can
be used for horizontal, uphill and downhill flow.

Weymouth:
This correlation is of the same form as the Panhandle and Modified Panhandle
correlations. It was originally developed for short pipelines and gathering systems. As a
result, it only accounts for horizontal flow and not for hydrostatic pressure drop.

School of Petroleum Engineering, UNSW


62

VirtuwellT M applies the standard hydrostatic head equation to account for the vertical
component of pressure drop. Thus, the implementation of the Weymouth equation
includes both horizontal and vertical flow components, and this equation can be used for
horizontal uphill and downhill flow.

In this software, for cases that involve a single-phase flow, the Gray, Hagedorn & Brown
and Beggs & Brill correlations revert to the Fanning single-phase correlations. For
example, if the Gray correlation was selected but there was only gas in the system, the
Fanning Gas correlation would be used. Similarly, for single-phase flow, the Flanigan
and Modified Flanigan correlations devolve to the single-phase Panhandle and Modified
Panhandle correlations respectively. The Weymouth (Multiphase) correlation devolves to
the single-phase Weymouth correlation.

Virtuwell TM Multiphase Flow Correlations:


Many of the published multiphase flow correlations are applicable for "vertical flow'
only, while others apply to "horizontal flow" only. Other than the Beggs and Brill
correlation, there are not many correlations that were developed for the whole spectrum
of flow situations that can be encountered in oil and gas operations; namely uphill,
downhill, horizontal, inclined and vertical flow. VirtuwellT M has adapted all of the
correlations (as appropriate) so that they apply to all flow situations. The following is a
list of the multiphase flow correlations that are available:

Gray:
The Gray Correlation (1978) was developed for vertical flow in wet gas wells.
FASTWELL uses a modified version of it so that it applies to flow in all directions by
calculating the hydrostatic pressure difference using only the vertical elevation of the
pipe segment and the friction pressure loss based on the total pipe length.

Hagedorn and Brown:


The Hagedorn and Brown Correlation (1964) was developed for vertical flow in wells.
FASTWELL uses a modified version of this correlation so that it applies to flow in all
directions by calculating the hydrostatic pressure difference using only the vertical
elevation of the pipe segment and the friction pressure loss based on the total pipe length.

Beggs and Brill: The Beggs and Brill Correlation (1973) is one of the few published
correlations capable of handling all of the flow directions. It was developed using
sections of pipe that could be inclined at any angle.

Flanigan:
The Flanigan Correlation (1958) is an extension of the Panhandle single-phase
correlation for multiphase flow. It incorporates a correction for multiphase Flow
Efficiency, and a calculation of hydrostatic pressure difference to account for uphill flow.
There is no hydrostatic pressure recovery for downhill flow. In this software, the
Flanigan multiphase correlation is also applied to the Modified Panhandle and Weymouth
correlations. It is recommended that this correlation not be used beyond +1- 10 degrees
from the horizontal.

School of Petroleum Engineering, UNSW


63

Modified-Flanigan:
The Modified-Flanigan is an extension of the Modified Panhandle single-phase equation
for multiphase flow. It incorporates the Flanigan correction of the Flow Efficiency for
multiphase flow and calculation of hydrostatic pressure difference to account for uphill
flow. There is no hydrostatic pressure recovery for downhill flow. In this software, the
Flanigan multiphase correlation is also applied to the Panhandle and Weymouth
correlations. It is recommended that this correlation not be used beyond +/- 10 degrees
for the horizontal.

Weymouth (Multiphase):
The Weymouth Correlation is an extension of the Weymouth single-phase equation for
multiphase flow. It incorporates the Flanigan correction of the Flow Efficiency for
multiphase flow and calculation of hydrostatic pressure difference to account for uphill
flow. There is no hydrostatic pressure recovery for downhill flow. In this software, the
Flanigan correlation is also applied to the Panhandle and Modified Panhandle
correlations. It is recommended that this correlation not be used beyond +/- 10 degree
from the horizontal.

Each of these correlations was developed for its own unique set of experimental
conditions, and accordingly results will vary between them.
In the case of single-phase gas, the available correlations are the Panhandle, Modified
Panhandle, Weymout and Fanning Gas. These correlations were developed for horizontal
pipes, but have been adapted to vertical and inclined flow by including the hydrostatic
pressure component. In vertical flow situations, the Fanning Gas calculation is equivalent
to a multi-step Cullender and Smith calculation.
In the case of single-phase liquid, the available correlation is the Fanning Liquid. It has
been implemented apply to horizontal, inclined and vertical wells.
For multiphase flow in essentially horizontal pipes, the available correlations are Beggs
& Brill, Gray, Hagedorn & Brown, Flanigan, Modified-Flanigan and Weymouth
(Multiphase). All of these correlations are accessible on the Pipe Module and the
Comparison Module of the program.
Warning: The Flanigan, Modified-Flanigan and Weymouth (Multiphase) correlations
can give erroneous results if the pipe described deviates substantially (more than 10
degrees) from the horizontal. The Gray and Hagedorn & Brown correlations were derived
for vertical wells and may not apply to horizontal pipes.
For multiphase flow in essentially vertical wells, the available correlations are Beggs &
Brill, Gray and Hagedorn & Brown. If used for single-phase flow, these three correlations
devolve to the Fanning Liquid correlation.
When switching from multiphase flow to single-phase flow, the correlation will default to
the Fanning. When switching from single-phase flow to multiphase flow, the correlation
will default to the Beggs and Brill.

School of Petroleum Engineering, UNSW


64

Example 2.6: Virtuwell TM

Use VirtuwellT M and the following data to answer the questions below:

ID OD
Tubing size
2.441" 2.875"
ID OD
Casing size 6.049" 7"
Total length of tubing 6500 ft
Total depth of hole 7090 ft
Perforated interval 6610 - 6625 ft
Oil production 4000 bbl/day
Water production 400 bbl/day
Gas production 0.5 MMcfd
Oil API gravity 35 API
Water SG 1.038
Gas SG 0.65
Bubble point of produced fluid 1200 psi
Reservoir pressure 4000 psi
Tubing head temperature 80o F
Bottomhole temperature 260o F

A) What is the bottomhole flowing pressure if the wellhead pressure is 900psi? Use the
following correlations:

i) Modiffied Beggs & Brill


ii) Gray
iii) Hagedorn &Brown

B) Plot the pressure profile for this case:

i) Change the gas production from 0.5 MMcfd to 5 MMcfd. What happens to
the plot?
ii) What correlations are most applicable for calculations on gas wells?

C) Plot the Oil IPR/TPR curve for this tubing (with gas production set back to 0.5
MMcfd).

i) Add IPR/TPR curves for the following tubing sizes:


a. OD 3.56, ID 3.5534
b. OD 4.5, ID 4.090
c. OD 5.5, ID 5.120
ii) Of these tubings, which one gives the highest production?

School of Petroleum Engineering, UNSW


65

Solution:

Part A:

First, input the data parameters as shown in the sample screen below. A schematic of the
wellbore will be drawn in the Schematic Module.

Properties of producing fluid should be entered in Fluid Properties tab.

In the Wellbore module, we can use different empirical correlations to determine the
sandface pressure given the wellhead pressure - or vice-versa. This particular screen
shows the results for the Modified Beggs and Brill Correlation.

School of Petroleum Engineering, UNSW


66

After placing the inputs in, we can select another correlation. The results for different
correlations are as follow:

Correlation Sandface Pressure (Psia)


Modified Beggs and Brill 3396.7
Gray 3281.6
Hagedorn and Brown 3347.1

Part B:

Increasing the gas production rate from 0.5 to 5.0 MMSCFD, the calculated sandface
pressure considerably falls. The results are summarized in following table:

Correlation Sandface Pressure (Psia) Sandface Pressure (Psia)


Gas Rate: 0.5 MMSCFD Gas Rate: 5 MMSCFD
Modified Beggs and Brill 3396.7 3182.7
Gray 3281.6 2670.4
Hagedorn and Brown 3347.1 2722.9

School of Petroleum Engineering, UNSW


67

VirtuwellT M can also be used to compare the pressure profile of fluids flowing through
specified tubing based on different correlations. To do this we use the Comparison
Module of the program (shown in VirtuwellT M screen below for the given case in this
example).

We can see from the results of this module that different correlations predict almost
similar pressure profiles. This is primarily attributed to the flow conditions and
composition. However, if we now change the gas production from 0.5 MMcfd to 5
MMcfd, we can see a significant change in the pressure profile. This is shown in the
VirtuwellT M screen below:

School of Petroleum Engineering, UNSW


68

We can see that there is a significant change in the pressure profile plots. Multiphase flow
in vertical wells can be best modeled by the Gray, Beggs & Brill, or Hagedorn & Brown
correlations.

School of Petroleum Engineering, UNSW


69

Part C
We can use the Oil IPR/TPC module in this software to generate tubing performance
curves (TPC curves) and productivity index curves (IPR curves). This module is shown
in the screen below:

From this module we can see that the rate of production will increase with the size of the
tubing. Hence, the largest tubing will have the highest production. However, we notice
that the incremental increase in production when using larger tubing is only marginal.
Therefore, we may wish to consider using the 3.56" tubing to complete this well. This
example illustrates how computers are used in the oil industry in the design and
simulation of tubing performance.

School of Petroleum Engineering, UNSW


70

REFERENCES

1. Allen, TO and Roberts, AP, Well Completion Design- Production Operations-1, 3rd
edition, 1989, pp 151-165.

2. Buzarde Jr, LE,Kastor, RL ,Bell, WT and DePriester,CL,1972,Well Completion


practices-Production Operations Course -1,1972, pp 45-63.

3. Brown, KE, Overview of Artificial Lift Systems. SPE 9979. SPE-AIME, 1982.

4. Breggs,HD and Brill,JP, Two phase flow in pipes, University of Tulsa, Tulsa, 1978 .

5. Dake, LP, Fundamentals of Reservoir Engineering. NY, Elsevier, 1978.

6. Elkins, LF, Skov, AM. and Liming, H.FA Practical Approach to Finding and
Correcting Perforation Inadequacies, Preprint of paper 2998 presented at 45 th Annual
Fall Meeting of SPE, Houston, Texas,1970.

7. Fetkovich, MJ, The Isochronal Testing of Oil Wells, SPE 4529. SPE-AIME, 1973.

8. Fetkovich, M.J, Multipoint Testing of Gas Wells, SPE Mid-continent section


Continuing Education Course of Well Test Analysis, March 17 1975.

9. Gilbert, WE, Flowing and Gas-Lift Well Performance, API Paper 801-30H, 1954.

10. Golan, M and Whitson, C, Well Performance, International Human Resources


Development Corporation, 1986.

11. Griffith and Wallis, GB, Two Phase Slug Flow, J, Heat transfer, ASME, Ser C, 83,
1961 pp 307-321.

12. Hagedorn, AR, and Brown, KE, Experimental study of pressure gradients occurring
during continuos two phase flow in small diameter conduits, JPT, 1965.

13. Muskat, M. and Evinger, H.H, Calculations of Theoretical Productivity Factor, Trans.
AIME, 1942,pp126-139.

14. Muskat, M, Physical Principles of Oil Production, McGraw-Hill Book Co., Inc., NY
1949, pp 210-214.

15. Rawlins, EL. and Schellhardt, MA., Back-Pressure Data on Natural Gas Wells and
Their Application to Production Practices, US Bureau of Mines Monograph 7,1936.

16. Standing, MB, Inflow Performance Relationships for Damaged Wells Producing by
Solution-Gas Drive, JPT, Nov. 1970, pp1399-1400.

School of Petroleum Engineering, UNSW


71

17. Standing, MB, Concerning the Calculation of Inflow Performance of Wells Producing
From Solution-Gas Drive Reservoirs, JPT, pp1141-1142.

18. Tarek Ahmed, Oil Well Performance, Gas Well performance-Reservoir Engineering
Handbook, 2000,pp 473 -568.

19. Vogel, JV, Inflow Performance Relationships for Solution-Gas Drive Wells, SPE
1476. SPE-AIME, 1966.

School of Petroleum Engineering, UNSW


72

REVIEW QUESTIONS

1. What is drawdown?

2. For oil reservoirs in an undersaturated region, how do we define the productivity


index (J)?

3. For gas reservoirs what is the exact solution for the Darcys equation for
compressible fluids under pseudo steady-state conditions?

4. What is the main difference between the Hagedorn & Brown correlation and the
Beggs & Brill correlation?

5. What is liquid holdup?

6. When can kinetic energy effects become significant on bottom hole flowing
pressure?

7. A 7000-ft well is to be produced with a target of 15,000 bbl/D. What tubing


intake pressure must be achieved to meet this target? Estimate tubing size and
make an educated guess as to whether artificial lift will be needed to produce
against a wellhead pressure of 400 psi. The well encountered 170 ft of oil-bearing
formation with pressure of 3000 psia. The hydrocarbon saturation is 80% and the
net-to-gross pay ratio is 50%. These development wells are being drilled with a
spacing of 3000 ft between wells (200-acre spacing). The production casing is 9
5 set in a 12 hole. The oil has a GOR of 500 scf/bbl, a formation volume
8

factor of 1.2, and a viscosity under reservoir conditions of 1.1 cp. Core tests have
shown that the average permeability to air is 435 mD and the average porosity
from both cores and log data is 28%. Experience has shown that the average well
can be expected to have a skin of +2. Relative permeability data are as follows:

Sw (%) k ro (%) k rw (%)


10 90 0
20 70 0
30 60 5
50 36 20
70 12 50
80 0 70

School of Petroleum Engineering, UNSW

Das könnte Ihnen auch gefallen