See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/378129680

DESIGN AND EVALUATION OF ESP AND GAS LIFT METHODS ON THE

OILFIELD WELLS

Technical Report · February 2024

DOI: 10.13140/RG.2.2.17685.70886

CITATIONS READS

0 194

1 author:

Elia Sikanyika
China University of Geosciences
10 PUBLICATIONS 4 CITATIONS

SEE PROFILE

All content following this page was uploaded by Elia Sikanyika on 11 February 2024.

The user has requested enhancement of the downloaded file.

CHINA UNIVERSITY OF GEOCIENCES
(WUHAN)

(中国地质大学（武汉）

FINAL COURSE PAPER PROJECT AND

HOMEWORK #01
Course: PETROLEUM PRODUCTION
ENGINEERING
TITLE: DESIGN AND EVALUATION OF ESP AND
GAS LIFT METHODS ON THE GYDA OILFIELD
WELLS

Lecturer: Dr. FENG YANG

Student Name: ELIA WILINASI SIKANYIKA

Reg Number: 1201990012

Major: M.Sc. PETROLEUM ENGINEERING

Date of Submission: 28, MAY 2020

TABLE OF CONTENTS
LIST OF FIGURES ............................................................................................................. iii

LIST OF TABLES ................................................................................................................ v

HOMEWORK #01 ................................................................................................................ 1

1. ABSTRACT .................................................................................................................. 6

2. INTRODUCTION ......................................................................................................... 7

3. THE GYDA OILFIELD................................................................................................. 8

3.1. Facts [23] ................................................................................................................ 9

3.2. Field description ................................................................................................... 10

4. OIL WELL PERFORMANCE ..................................................................................... 12

4.1. Drive mechanisms ................................................................................................. 12

4.2. Inflow performance ............................................................................................... 13

4.3. Outflow performance ............................................................................................ 16

4.4. Operating point ..................................................................................................... 19

5. MULTIPHASE FLOW ................................................................................................ 20

6. ARTIFICIAL LIFT ...................................................................................................... 23

6.1. Artificial lift on Gyda oilfield wells ....................................................................... 24

6.2. Artificial lift selection ........................................................................................... 25

6.2.1. Selection by Consideration of Depth/Rate System .......................................... 25

6.2.2. Selection by Advantages and Disadvantages .................................................. 26

6.2.3. Selection by Net Present Value Comparison .................................................. 29

7. ELECTRIC SUBMERSIBLE PUMPS (ESPS) ............................................................. 30

7.1. Centrifugal pump .................................................................................................. 33

7.2. Seal Chamber section ............................................................................................ 37

7.3. The Motor ............................................................................................................. 37

7.4. The power cable .................................................................................................... 38

7.5. ESP run life ........................................................................................................... 39

i|P age
 7.5.1. Case studies ................................................................................................... 43

7.6. ESP design for Gyda ............................................................................................. 45

7.7. ESP secondary effect ............................................................................................ 47

8. GAS LIFT .................................................................................................................... 48

8.1. The Unloading Process ......................................................................................... 50

8.2. Gas lift performance curve .................................................................................... 52

8.3. Gas lift valves ....................................................................................................... 52

8.4. Gas lift completion procedure ............................................................................... 57

9. PROSPER .................................................................................................................... 58

9.1. Basic theory of PROSPER .................................................................................... 60

9.2. Building a base model for A-19 and A-26 ............................................................. 61

9.2.1. PVT ............................................................................................................... 62

9.2.2. IPR ................................................................................................................ 63

9.2.3. Equipment ..................................................................................................... 65

9.2.4. Results ........................................................................................................... 69

10. GAS LIFT DESIGN ................................................................................................. 71

10.1. Modelling A-19 and A-26 with gas lift .............................................................. 72

10.2. Positioning of valves ......................................................................................... 75

10.3. Results............................................................................................................... 77

10.4. Sensitivities of injection depth ........................................................................... 77

11. ESP DESIGN ........................................................................................................... 79

11.1. Modelling A-19 and A-26 with ESP .................................................................. 81

11.2. Results............................................................................................................... 83

12. PRODUCTION FORECAST ................................................................................... 84

12.1. Well A-19.......................................................................................................... 85

12.2. Well A-26.......................................................................................................... 87

13. ECONOMICAL EVALUATION ............................................................................. 89

ii | P a g e
14. CONCLUSIONS ...................................................................................................... 90

REFERENCES ................................................................................................................... 92

ABBREVIATIONS ............................................................................................................ 93

LIST OF FIGURES
Figure 1; IPR curves for Well A. ........................................................................................... 3
Figure 2;IPR curves for Well B. ............................................................................................ 5
Figure 3; Southern North Sea. ............................................................................................... 8
Figure 4; The Gyda platform. ................................................................................................ 9
Figure 5; Gyda reservoir. ..................................................................................................... 11
Figure 6; Production history of the Gyda field. .................................................................... 12
Figure 7; IPR curve. ............................................................................................................ 15
Figure 8; VLP curve ............................................................................................................ 19
Figure 9; Operating point (intersection between IPR and VLP curves). ............................... 19
Figure 10; Flow regimes in horizontal flow [5]. ................................................................... 20
Figure 11; Flow regimes in vertical flow [5]. ....................................................................... 21
Figure 12; Flow regime map for vertical flow [5]. ............................................................... 22
Figure 13; Run life of ESP systems. The figure is based on Centrilift data. [13] .................. 28
Figure 14; Basic ESP .......................................................................................................... 31
Figure 15; ESP surface system. ........................................................................................... 32
Figure 16; the inside of a centrifugal pump. ......................................................................... 33
Figure 17; Shaft with the rotating impellers attached ........................................................... 34
Figure 18; ESP operating range ........................................................................................... 35
Figure 19; Standard pump curves for head, efficiency and BHP [19]. .................................. 36
Figure 20; Seal. ................................................................................................................... 37
Figure 21; round design....................................................................................................... 38
Figure 22; flat design .......................................................................................................... 38
Figure 23; Factors acting on ESP run life [13] ..................................................................... 39
Figure 24; ESP run life on Beatrice. The figure shows data from two HP intervals [13]. ...... 44
Figure 25; Dual ESP system design for Gyda ...................................................................... 46
Figure 26; ADV operating schematic [22] ........................................................................... 47

iii | P a g e
Figure 27; ESP secondary effect. Pressure drawdown makes oil migrate into the high perm.
Zone.................................................................................................................................... 48
Figure 28; Gas lift. Gas is pumped down the annulus and into the tubing [18]. .................... 49
Figure 29; the unloading process. ........................................................................................ 51
Figure 30; Gas lift performance curve. The figure shows the total production rate plotted
against the rate of lift gas injected [14]. ............................................................................... 52
Figure 31; conventional unloading valve [24]. ..................................................................... 53
Figure 32; IPO gas lift valve operating principle [2] ............................................................ 55
Figure 33; cross section of square edge orifice and venturi valve [2]. .................................. 56
Figure 34; Gas passage characteristics comparison [2]. ....................................................... 56
Figure 35; Valve installation with KOT. [24] ...................................................................... 58
Figure 36; PROSPER front display. .................................................................................... 59
Figure 37; System summary. ............................................................................................... 62
Figure 38; PVT correlations. ............................................................................................... 63
Figure 39; survey data for A-19. The well path is plotted to the right. .................................. 65
Figure 40; survey data for A-26. The well path is plotted to the right. .................................. 66
Figure 41; Completion schematic of A-19 ........................................................................... 67
Figure 42; Completion schematic of A-26 ........................................................................... 68
Figure 43; production point of A-19 in May 2010 ............................................................... 69
Figure 44; production point of A-26 in May 2010 ............................................................... 70
Figure 45; A-26 with no intersection between IPR and VLP (situation today). .................... 70
Figure 46; gas lift design menu............................................................................................ 72
Figure 47; Gas lift performance curve for A-19 ................................................................... 73
Figure 48; Gas lift performance curve for A-26 ................................................................... 74
Figure 49; valve spacing for A-19 ....................................................................................... 75
Figure 50; valve spacing for A-26 ....................................................................................... 76
Figure 51; injection depth analysis on A-19. The figure shows how the VLP curve is moved
as a function of injection depth. ........................................................................................... 78
Figure 52; injection depth analysis on A-26. The figure shows how the VLP curve is moved
as a function of injection depth. ........................................................................................... 79
Figure 53; Centrilift 562 P110 [19]...................................................................................... 81
Figure 54; pump curves for the Centrilift 562 P110 ............................................................. 82
Figure 55; ESP input in PROSPER (A-19 as an example) ................................................... 83
Figure 56; A-19 ESP system ............................................................................................... 84

iv | P a g e
Figure 57; Production forecast for A-19. Production for the base case and the two artificial
lift methods are plotted from start date to 2019. ................................................................... 86
Figure 58; production forecast for A-26. Production for the base case and the two artificial lift
methods are plotted from start date to 2019. ........................................................................ 88

LIST OF TABLES

Table 1; show calculated points for pfw (psi) and q (stb/day) for Well A ............................... 2
Table 2; show calculated points for pfw (psi) and Q (stb/day) for Well B .............................. 4
Table 3; Advantages and disadvantages of ESP and gas lift [4] ........................................... 26
Table 4: Design considerations and overall Comparisons [7] ............................................... 27
Table 5; Wytch Farm ESP run life [13] ............................................................................... 45
Table 6; General PVT input for A-19 and A-26 ................................................................... 63
Table 7; IPR input for A-19................................................................................................. 64
Table 8; IPR input for A-26................................................................................................. 65
Table 9; Temperature profile ............................................................................................... 69
Table 10; Results from gas lift design.................................................................................. 77
Table 11; injection depth analysis on A-19 and A-26........................................................... 78
Table 12; ESP system data [19] ........................................................................................... 80
Table 13; Results from ESP simulation ............................................................................... 83
Table 14; A-19 “base case” ................................................................................................. 85
Table 15; A-19 “ESP case” ................................................................................................. 86
Table 16; A-19 “gas lift case” ............................................................................................. 86
Table 17; A-26 “ESP case” ................................................................................................. 87
Table 18; A-26 “gas lift case” ............................................................................................. 87
Table 19; capital cost of projects. ........................................................................................ 89

v|P ag e
 CHAPTER 01
HOMEWORK #01
QUESTION

Construct IPR of two wells in an undersaturated oil reservoir using the generalized
Vogel equation. The following data are given:

Reservoir pressure: p̅ = 5,000 psi

Bubble point pressure: pb = 3,000 psi

Tested flowing bottom-hole pressure in Well A: pwf1 = 4,000 psi

Tested production rate from Well A: q1 = 300 stb/day

Tested flowing bottom hole pressure in Well B: pwf1 = 2,000 psia

Tested production rate from Well B: q1 = 900 stb/day

SOLUTION:

FOR WELL A:

Data Given:

Reservoir pressure, p̅ = 5,000 psi

Bubble point pressure, pb = 3,000 psi

Tested flowing bottom-hole pressure in Well A, pwf1 = 4,000 psi

Tested production rate from Well A, q1 = 300 stb/day

pwf1 (4,000 psia) > pb (3,000 psi)

1|P a ge
Since bottom-hole pressure, pwf1 (4,000 psia) is greater than bubble-point
pressure, pb (3,000 psi) the model constant J* should be determined using equation
below;

q
The productivity index, J ∗ = ( p̅ −p1
wf1 )

300 stb/day
J ∗ = ( 5000 psi −4000 psi)

The productivity index, 𝐉 ∗ = 0.3000 sb/day-psi

Calculation of points used for drawing an IPR plot for Well A

The generalized Vogel equation below has used to calculate points for pfw (psi) and q
(stb/day) for Well A and Calculated points have shown in table 1 below;

𝐉 ∗𝐩𝐛 𝐩 𝐩 𝟐
q = 𝐉 ∗ (𝐩
̅ − 𝐩𝐛 ) + × [𝟏 − 𝟎. 𝟐 ( 𝐩𝐰𝐟 ) − 𝟎. 𝟖 ( 𝐩𝐰𝐟 ) ]
𝟏.𝟖 𝐛 𝐛

Table 1; show calculated points for pfw (psi) and q (stb/day) for Well A

pwf (psi) q (stb/day)

0 1,100
500 1,072
1,000 1,022
1,500 950
2,000 856
2,500 739
3,000 600
5,000 0

The IPR curve is plotted in Fig. 1 below

2|P a ge
 IPR Curve for Well A
6000

5000

4000

pwf, psia 3000

2000

1000

0
0 200 400 600 800 1,000 1,200
q, stb/day

Figure 1; IPR curves for Well A.

FOR WELL B:

Data Given:

Reservoir pressure, p̅ = 5,000 psi

Bubble point pressure, pb = 3,000 psi

Tested flowing bottom hole pressure in Well B, pwf1 = 2,000 psia

Tested production rate from Well B, q1 = 900 stb/day

pwf1 (2,000 psia) < (pb 3,000 psi)

Since bottom-hole pressure, pwf1 (2,000 psia) is less than bubble-point

pressure, pb (3,000 psi) the model constant J* should be determined using equation
below;

q1
The productivity index, J ∗ = pb p p 2
̅ −pb )+ [1−0.2( wf1 )−0.8( wf1 ) ])
(( p
1.8 pb pb

900 stb/day
J∗= 3,000 psi 2,000 psia 2,000 psia 2
(( 5000 −3,000 psi)+ [1−0.2( )−0.8( ) ])
1.8 3,000 psi 3,000 psi

3|P a ge
The productivity index, 𝐉 ∗ = 0.3156 stb/day-psi

Calculation of points used for drawing an IPR plot for Well B

The generalized Vogel equation below has used to calculate points for pfw (psi) and q
(stb/day) for Well B and Calculated points have shown in table 2 below;

𝐉 ∗𝐩𝐛 𝐩 𝐩 𝟐
q = 𝐉 ∗ (𝐩
̅ − 𝐩𝐛 ) + × [𝟏 − 𝟎. 𝟐 ( 𝐩𝐰𝐟 ) − 𝟎. 𝟖 ( 𝐩𝐰𝐟 ) ]
𝟏.𝟖 𝐛 𝐛

Table 2; show calculated points for pfw (psi) and Q (stb/day) for Well B

pwf (psi) q (stb/day)

0 1,157
500 1,128
1,000 1,075
1,500 999
2,000 900
2,500 777
3,000 631
5,000 0

The IPR curve is plotted in Fig. 2 below

4|P a ge
 IPR Curve for Well B
6000

5000

pwf, psia 4000

3000

2000

1000

0
0 200 400 600 800 1,000 1,200 1,400
q, stb/day

Figure 2; IPR curves for Well B.

5|P a ge
 CHAPTER 02

DESIGN AND EVALUATION OF ESP AND GAS LIFT

METHODS ON THE GYDA OILFIELD WELLS
*ELIA WILINASI SIKANYIKA*

1. ABSTRACT
Since a production peak in 1995 the oil production on the Gyda field has decreased. Water cut
is increasing and reservoir pressure is decreasing. This final course paper report is a study of
the artificial lift methods being evaluated to increase the production in the late life of the field.

A thorough investigation of gas lift and Electrical Submersible Pump (ESP) theory, design, and
production output is carried out. The theory of artificial lift selection is also presented.

Based on reservoir inputs and completion design, production has been simulated in PROSPER
for different scenarios and methods.

The main conclusions and recommendations are as follows:

 Gas lifting is a simple, well tried method where little can go wrong, while ESPs are a
complex solution which will require a large amount of planning and administrative
resources.
 ESPs have a limited lifetime which increases cost later in a project. The expected
lifetime of an ESP well on Gyda is two years. The initial cost for a gas lift well and ESP
well will not be so different, because a lot of the wells need a full workover before they
can be used for gas lifting.
 Production through gas lifting is not only dependent on injection rate, but can be
optimized through the completion design. Setting the valves deeper gives an increased
production.
 A new gas compressor is needed if a gas lifting campaign is to be initiated.

6|P a ge
  Baker Hughes Centrilift’s ESP design was verified for the start-up (May 2010)
conditions. But production can fall beneath minimum design rate after some years, and
a new evaluation of design must be done when the pumps are changed at failure.
 The production simulation of the pilot wells A-19 and A-26 shows that the ESP solution
is superior to the gas lift. A secondary effect from the ESP pressure drawdown can also
increase production and recovery factor for the field.
 Even though ESPs seem to be the superior choice, an economical evaluation of the
projects entire lifetime is needed. A Net Present Value analysis will give the different
projects a comparable value, which includes the costs and production from start to
finish.

2. INTRODUCTION
The Talisman Energy Norge operated Gyda field is in its late life production. The water
production is increasing and the reservoir pressure is decreasing. To increase production and
extend the lifetime of the field, the operator has decided start an artificial lift project. Because
of conditions and desired rates gas lifting and Electrical Submersible Pumps (ESPs) has been
evaluated to be the only two alternatives.

The operator is looking at either a full field (10 wells) gas lift campaign or ESPs. If the ESP
solution would be found to be the best method, it has been decided that a pilot project is
necessary to see how the field responds to the pumps. For this well A-19 and A-26 has been
chosen because of their production potential.

The objective of this study is to make a decision supporting document, evaluating each of the
scenarios in detail. Simulations based on the field data will give an indication of what rates the
different solutions will give.

The first part of the report paper describes the basic theory of well performance and physics.
Second, a detailed description of gas lift and ESP methodology and design is given. Theory on
decision making within artificial lift is also presented.

The third part of the study is a practical simulation. PROSPER is chosen as the tool to simulate
production in the pilot wells A-19 and A-26. Three models are built for each well presenting a
“base case”, “ESP case” and “gas lift case”.

7|P a ge
For gas lift design the simulation tool is used, but the ESP design is provided by Baker Hughes
Centrilift. Based on reservoir simulations a production forecast is made until 2019 for each
scenario.

The last section gives a short economical evaluation of the projects, which is an important part
of a final decision on artificial lift method.

3. THE GYDA OILFIELD

The Gyda field is located southwest in the North Sea, 270 km from Stavanger and 43 km
northwest of Ekofisk (Fig. 3).

Figure 3; Southern North Sea.

It has been developed with a combined drilling, accommodation and processing facility with a
steel jacket. The sea depth in the area is 66 metres. The oil is transported to Ekofisk via the oil

8|P a ge
pipeline from Ula and in Norpipe to Teeside. The gas is transported in a dedicated pipeline to
Ekofisk for onward transport in Norpipe to Emden [23].

Figure 4; The Gyda platform.

3.1. Facts [23]

Found: The field was discovered by exploration well 2/1-3 in 1980

Production start: 21.06.1990

Operator: Talisman Energy Norge AS (Since 2003)

Owners: DONG E&P Norge AS 34, 00 %

Norske AEDC A/S 5, 00 %

Talisman Energy Norge AS 61, 00 %

Recoverable reserves: Remaining reserves 31.12.2008:

9|P a ge
38,8 million Sm3 oil 4,1 million Sm3 oil

6,2 billion Sm3 gas 0,3 billion Sm3 gas

1,9 million-ton NGL 0,0 million-ton NGL

3.2. Field description

The Gyda reservoir consists of Upper Jurassic shallow marine sandstone. The trap combines
both structural and stratigraphic elements. The field is broadly a westward dipping and
westward thickening wedge. The reservoir lies between 3680 and 4170 m TVDSS. The
overlying Kimmeridgian aged Mandal formation provides both the seal and the hydrocarbon
source for the Gyda Field.

The sands are divided into 3 main units A-, B- and C-sand. The A-sand is in the bottom with a
high permeability zone at the top. The permeability in the top of the A- sand is up to 1 D while
the base can be 1 mD and below. The B-sand is the middle sand, and in general has poor
reservoir quality. The best parts of the sand have permeabilities around 30 mD, while most of
it is around 1 mD. The C-sand on the top pinches out towards the crest of the field, and varies
in reservoir quality. The C-sand is interbedded with calcite stringers and the eastern parts have
poor reservoir quality, equivalent to the B-sand. The western parts of the C-sand have very
good reservoir quality, up to 800 mD in places.

The reservoir is cut by numerous Late Jurassic faults with variable throws. Several studies
suggest that faults and fractures are at least initially sealing. This creates the opportunity for
compartmentalization within the reservoir.

The field is divided into three main segments: Main, South West and South. These segments
are confirmed by differences in reservoir fluids, original oil water contacts (OWC) and
dynamic pressure data.

Of the 32 wells on Gyda, 17 are currently active. 11 are producing and 6 are injecting. The rest
is either temporarily closed or plugged and abandoned.

10 | P a g e
 Figure 5; Gyda reservoir.

Today Gyda is in its tail phase and experiences increasing water production (Fig. 6) and
challenges in maintaining the oil production. The production license period was recently
extended to 2028. Several new wells are being drilled on the field. A compressor was installed
in 2007 for a gas lift pilot. This has resulted in improved production from the wells. It is also
considered to tie-in other deposits in the area to Gyda.

11 | P a g e
 Figure 6; Production history of the Gyda field.

4. OIL WELL PERFORMANCE

Oil well performance is dependent on a large number of variables like pressures, formation
properties and fluid properties. And these are again dependent on each other.

Different models for the well inflow performance and the vertical lift performance will be
described in this chapter, but first a short explanation of the different drive mechanisms from
the reservoir will be given.

4.1. Drive mechanisms

According to Dake [3] oil production is due to the following drive mechanisms:

 Natural water drive

 Solution gas drive
 Gas-cap drive
 Compaction drive

12 | P a g e
Natural water drive

A drop in the reservoir pressure, due to the production of fluids, causes the aquifer water to
expand and flow into the reservoir. 50% of oil recovery can be caused by natural water drive.

Solution gas drive

When the reservoir pressure drops below the bubble point pressure solution gas dissolved in
oil appears as a free phase. When pressure drops further the highly compressible gas expands
expelling the oil from porous media.

Gas-cap drive

High gas compressibility and the extended gas cap size ensure a long lasting and efficient field
performance. Up to 35% of the original oil in place can be recovered under a gas-cap drive.

Compaction drive

This drive mechanism might occur during depletion when rock grains are subjected to stress
beyond elasticity limit. It leads to a re-compaction of partially deformed or even destroyed rock
grains that might result in gradual or abrupt reduction of the reservoir pore volume.

In order to achieve better field performance, secondary and tertiary oil recovery methods are
often implemented. Gas lift and downhole pumps are examples of advanced recovery
techniques (Enhanced Oil Recovery, EOR).

4.2. Inflow performance [1], [2]

The Inflow Performance Relationship (IPR) describes pressure drawdown as a function of
production rate, where drawdown is defined as the difference between static and flowing
bottom hole pressure (FBHP).

The simplest approach to describe the inflow performance of oil wells is the use of the
productivity index (PI) concept. It was developed using the following assumptions:

13 | P a g e
  Flow is radial around the well
 A single-phase liquid is flowing
 Permeability distribution in the formation is homogeneous
 The formation is fully saturated with the given liquid.

The flow through a porous media is given by the Darcy equation:

q k dp
=μ (Eq. 4.1)
A dl

Using the assumptions above it can be written as

0.00708kh
q= r (pR − pwf ) (Eq. 4.2)
μB ln( e )
rw

Where: q = liquid rate, STB/d

k = effective permeability, mD

h = pay thickness, ft

µ = liquid viscosity, cP

B = liquid volume factor, bbl/STB

re = drainage radius of well, ft

rw = radius of wellbore, ft

pR = average reservoir pressure

pwf = flowing bottomhole pressure

Most parameters on the right-hand side are constant, which permits collecting them into a
single coefficient called Production Index, PI:

0.00708kh
So, PI = r
μB ln( e )
rw

q = PI(pR − pwf ) (Eq. 4.3)

This gives us:

14 | P a g e
 q
PI = (p (Eq. 4.4)
R −pwf )

This equation states that liquid inflow into a well is directly proportional to the pressure
drawdown. It will plot as a straight line on a pressure vs. rate diagram.

The use of the PI concept is quite straightforward. If the average reservoir pressure and the PI
are known, use of equation 4.3 gives the flow rate for any FBHP. The well’s PI can either be
calculated from reservoir parameters, or measured by taking flow rates at various FBHPs.

This works well for a single-phase flow, but when producing a multiphase reservoir, the curve
will not plot as a straight line.

As the oil approaches the well bore and the pressure drops below bubble point, gas comes out
of solution. Thus, the free gas saturation in the vicinity of the oil steadily increases, which
implies that the relative permeability to gas steadily increases at the expense of the relative
permeability of oil. The greater the drawdown, the bigger this effect would be. Since the PI
depends on the effective oil permeability, it is expected that it will decrease (Eq. 4.2). Figure 7
shows the IPR curve for this condition.

Figure 7; IPR curve.

Vogel [9] used a numerical reservoir simulator to study the inflow of wells depleting solution
gas drive reservoirs. He considered cases below bubble point and varied parameters like draw

15 | P a g e
downs, fluid and rock properties. Vogel found that the calculated IPR curves exhibited the
same general shape, which is given by the dimensionless equation:

q p p 2
= 1 − 0.2 ( pwf ) − 0.8 ( pwf ) (Eq. 4.5)
qmax R R

The equation is generally accepted for other drive mechanisms as well, and is found to give
reliable results for almost any well with a bottom hole pressure below bubble point of the oil.

There are a number of other models designed for special cases e.g. horizontal wells, transient
flow, fractured wells, non-Darcy pressure loss, high rates etc.

4.3. Outflow performance

The well’s outflow performance, or Vertical Lift Performance (VLP), describes the bottomhole
pressure as a function of flow rates. According to Golan and Whitson [1] the outflow
performance is dependent on different factors; liquid rate, fluid type (gas- to-liquid ratio, water
cut), fluid properties and tubing size.

Gabor [2] divides the total pressure drop in a well into a hydrostatic component, friction
component and an acceleration component:

Hydrostatic component represents the change in potential energy due to gravitational force
acting on the mixture:

dp
( dl ) = ρgsinβ (Eq. 4.6)
h

Where: ρ = density of fluid

β = pipe inclination angle, measured from horizontal

g = gravity constant

Friction component stands for the irreversible pressure losses occurring in the pipe due to
fluid friction on the pipe inner wall:

16 | P a g e
 dp 1 1
( dl ) = d f2 ρv 2 (Eq. 4.7)
f

Where:

f = friction factor

d = pipe inside diameter

v = fluid velocity

The type of flow is determined from the Reynolds number:

ρvd
Re = (Eq. 4.8)
μ

Where: µ = fluid viscosity

The boundary between flow regimes are:

Re ≤ 2000: Laminar flow

2000 < Re ≤ 4000: Transition between laminar and turbulent flow

4000 < Re: Turbulent flow

For laminar flow f = 64/Re (Moody friction factor). However, finding the friction factor is
more complicated for turbulent flow, and there are several ways to calculate the friction factor.

2
1
Darcy friction factor, f (D) = ( Relative roughness 1.11 6.9
)
−1.8×log10 [( ) + ]
3.7 NR

But also;
f (D)
Fanning Friction factor, F = 4

Acceleration component represents the kinetic energy changes of the flowing mixture and is
proportional to the changes in flow velocity. The term is often negligible:

17 | P a g e
 dp dv
( dl ) = −ρv dl (Eq. 4.9)
a

Other Effects

Effect of liquid flow rate on pressure loss

From the friction equation we can see that friction losses increases as liquid rate increases (v
increases). Hydrostatic gradient also increases with increased liquid production.

Effect of gas-to-liquid ratio on pressure loss

Increase in gas-to-liquid ratio (GLR) results in reduction of hydrostatic gradient. On the other
hand, increased GLR increases friction forces and has a counter effect on the bottomhole
pressure. When contribution of the friction becomes higher than that of hydrostatic forces, the
actual bottomhole pressure starts to increase. From a gas lift point of view this means that there
is a limit of how much gas that beneficially can be injected.

Effect of water cut on pressure loss

Increased water cuts result in increased liquid density, which in turn, increases hydrostatic
forces and the bottomhole pressure

Effect of tubing size on pressure loss

From the equation 4.7 we can see that the increased diameter of tubing reduces the pressure
gradient due to friction. However, there is a limit to which diameter of tubing can be increased.
If the diameter is too big the velocity of the mixture (v = q/A, A: pipe cross section) is not
enough to lift the liquid and the well starts to load up with liquid, resulting in increase of
hydrostatic pressure.

18 | P a g e
 Figure 8; VLP curve

4.4. Operating point

To calculate the well production rate, the bottom-hole pressure that simultaneously satisfies
both the IPR and VLP relations is required. By plotting the IPR and VLP in the same graph the
producing rate can be found. The system can be described by an energy balance expression,
simply the principle of conservation of energy over an incremental length element of tubing.
The energy entering the system by the flowing fluid must equal the energy leaving the system
plus the energy exchanged between the fluid and its surroundings.

Figure 9; Operating point (intersection between IPR and VLP curves).

19 | P a g e
 5. MULTIPHASE FLOW
Oil wells normally produce a mixture of fluids and gases to the surface while phase conditions
usually change along the path. At higher pressures, especially at the well bottom, flow may be
single phase. But going up in the well the continuous decrease of pressure causes dissolved gas
to gradually escape from the flowing liquid, resulting in multiphase flow. Gas injection into a
well is also an example of multiphase flow.

In single phase flow we discriminate between laminar and turbulent flow. In two-phase flow
we discriminate in addition between flow regimes that are characteristic for the time and space
distribution of gas and liquid flow.

In horizontal flow we discriminate between the flow regimes

 Stratified flow
 Slug flow
 Dispersed bubble flow
 Annular flow
These are shown in figure 10. At low velocities the gas and liquid are separated as in stratified
flow. At high velocities gas and liquid become mixed. Slug flow is an example of a flow regime
in between, representing both separation and mixing. Slug flow is consequently referred to as
an intermittent flow regime [5].

Figure 10; Flow regimes in horizontal flow [5].

20 | P a g e
In vertical flow we discriminate between the flow regimes

 Slug flow
 Churn flow
 Dispersed bubble flow
 Annular flow

Figure 11 illustrates the flow regimes in vertical flow. The same comments that apply to
horizontal flow are valid in vertical flow. The big difference is that in vertical (concurrent
upward) flow it is not possible to obtain stratified flow. The equivalent flow regime at identical
flow rates of gas and liquid is slug flow with very slow bullet shaped Taylor bubbles.

Figure 11; Flow regimes in vertical flow [5].

21 | P a g e
 Figure 12; Flow regime map for vertical flow [5].

The superficial velocities are defined by:

qL
ULS = (Eq. 5.1)
A

qG
UGS = (Eq. 5.2)
A

They are also referred to as apparent velocities or volumetric fluxes. From the definition we
see that the volumetric flow rates and the pipe cross section A is known, the superficial
velocities follow directly.

The phase velocities are the real velocities of the flowing phases. They may be defined locally
(at a certain position in the pipe cross section) or as a cross sectional average for the pipe. They
are defined by:
q
uL = AL (Eq. 5.3)
L

q
uG = AG (Eq. 5.4)
G

22 | P a g e
Gas and liquid in general flow with different phase velocities in pipe flow. The relative phase
velocity or the slip velocity is defined by:

u s = |u G − u L | (Eq. 5.5)

The slip velocity thus has the same unit as the phase velocities. In addition, the slip ratio is
commonly used:
uG
S= (Eq. 5.6)
uL

Note that the slip ratio is dimensionless. Slip effect is seen in inclined flow and is caused by
the density difference between the gas and liquid, which in turn causes a velocity difference;
the gas will rise through the liquid [5].

“Hold up” is a consequence of slip and is defined as the proportion of the pipe that is occupied
by liquid.

Multiphase flow correlations are used to predict the liquid holdup and frictional pressure
gradient. Correlations in common consider the oil and gas lumped together as one equivalent
fluid. They are therefore more correctly termed 2-phase flow correlations. Depending on the
particular correlation, flow regimes are identified and specialized holdup and friction gradient
calculations are applied for each flow regime. Some of the correlations most widely accepted
for oil wells are:

 Duns and Ros

 Hagedorn and Brown
 Orkiszewski
 Beggs and Brill

6. ARTIFICIAL LIFT
Maximizing the use of natural energy in a reservoir is critical to any production installation. In
a naturally flowing well there is enough energy stored in the reservoir to flow the produced
fluid to the surface. Reservoir pressure and formation gas provide this energy in the flowing
well. When reservoir energy is too low for natural flow, or when the desired production rate is
greater than the reservoir energy can deliver, it becomes necessary to put the well on some

23 | P a g e
form of artificial lift. As of 2006, 90 % of the world’s oil wells are on some form of artificial
lift according to Oilfield Review [16].

An oil well usually flows naturally initially, that means the pressure at well bottom is sufficient
to overcome the pressure losses in the well and flow line to the separator. When the criteria is
no longer met due to decrease in bottom hole pressure, or pressure losses in the well become
to great, the natural flow stops and the well dies. The increased pressure losses in the well can
come from increased overall density due to decreased gas production, increased water cut or
mechanical problems like downhole restrictions (scale etc.).

Artificial lift methods fall into two groups, those that use pumps and those that use gas [17].

Artificial Lift Types:

 Beam Pump / Sucker Rod Pumps (Rod Lift)

 Progressive Cavity Pumps (Jet /piston lift)
 Subsurface Hydraulic Pumps
 Electric Submersible Pumps (ESP) Gas Method:
 Gas Lift

6.1. Artificial lift on Gyda oilfield wells

Since a peak in 1995, production on Gyda has decreased (Fig. 6). This is due to a decreasing
reservoir pressure and increasing water cut which comes from a large amount of water being
injected to support the reservoir pressure.

A jet pump was tried on well A-26 in 1995 when BP was operator of the field [12]. The trial
was without success due to the plugging of the pump. Scale (calcium sulphate) was found in
and above the pump. During this study a second attempt on investigating the use of jet pumps
was done. The results showed that the solutions available had too little production potential
and the risk of the pumps scaling up again was too high.

In 1997 the use of ESPs was excluded from Gyda because they couldn’t withstand the high
downhole temperatures of 160 °C. However, the use of ESPs is now up for debate again as
pumps are made to run under tougher conditions. Technology developed from steam-assisted
gravity drainage (SAGD) operations is planned to be implemented on Gyda. The SAGD
equipment can withstand bottom hole temperatures of up to 218 °C [9].

24 | P a g e
During summer 2007 gas lift was installed on Gyda and has been a big success. The first well
to be subject for gas lift (A-17A) increased production rate from 400-500 bbl/d to 1100 bbl/d.
This was when all available gas was injected to this single well (56 600 Sm3/day). The second
well of interest (A-02A) died due to high water cut and low reservoir pressure. With gas lift
Talisman Energy Norge AS (TENAS) succeeded getting this well back on production with
initial rate of 1500 bbl/d. Today A-27A is also on gas lift.

As the production rate is decreasing there is no doubt that TENAS should continue improving
artificial lift on Gyda. The question is if gas lift or ESP is the best solution.

A-19 and A-26 has been chosen to be a part of a pilot project for artificial lift because of their
great potential of higher production. If the pilot project works out well, a full field artificial lift
campaign can be realized. Later in this report production is simulated for different scenarios in
A-19 and A-26.

6.2. Artificial lift selection

To realize the maximum potential from developing any oil or gas field, the most economical
artificial lift method must be selected. This part discusses some of the most commonly used
methods for selecting an artificial lift system. In most cases, what has worked best or which
lift method performs best in similar fields serve as selection criteria. Also, the equipment and
services available from vendors can easily determine which lift method will be applied.
However, when significant costs for well servicing and high production rates are a part of the
scenario, it becomes prudent for the operator to consider most, if not all, of the available
evaluation and selection methods.

If the “best” lift method is not selected, such factors as long-term servicing costs, deferred
production during workovers, and excessive energy costs (poor efficiency) can drastically
reduce the net present value (NPV) of the project.

6.2.1.Selection by Consideration of Depth/Rate System

One simple selection or elimination method is the use of charts that show the range of depth
and rate in which particular lift types can function. Charts like this are approximate for initial
selection possibilities along with advantage/disadvantage lists (see next section). Particular
well conditions, such as high viscosity or sand production, may lead to the selection of a lift

25 | P a g e
method not initially indicated by the charts. Specific designs are recommended for specific
well conditions to more accurately determine the rates possible from given depths [4].

6.2.2.Selection by Advantages and Disadvantages

Table 3 provides a useful summary of advantages and disadvantages of the two alternative
lifting systems on Gyda, while Table 4 gives an overall consideration. Much of the selection
process can be accomplished with depth rate charts and this extensive set of tables of artificial
lift capabilities. But very severe and special conditions can require further study.

Consideration of reservoir characteristics and location are examples of what will fall inn to this
category. If the well may be expected to decline rapidly, it would not be wise to choose a high-
volume method that will only be required for a short time.

Another example would be if there is a lack of electric power or economically supplied electric
power; the use of ESPs is not possible [8].

Table 3; Advantages and disadvantages of ESP and gas lift [4]

Advantages and disadvantages of ESP and gas lift [4]

ESP Gas lift

Advantage Disadvantage Advantage Disadvantage
Can lift extremely Only applicable with Can handle large Lift gas is not always
high volumes. electric power. volumes of solids with available.
minor problems.
Unobtrusive in urban High voltages (1000 Handles large Not efficient in lifting
locations V) are necessary. volume in high-PI small fields or one-
wells. well leases.
Applicable offshore Impractical in shallow, Unobtrusive in urban Difficult to lift
low volume locations. emulsions and
wells. viscous crude.
Corrosion and scale Expensive to change Power source can be Gas freezing and
treatment easy to equipment to match remotely located. hydrate problems.
perform. declining well
capability.
Simple to operate. Cable causes problems Lifting gassy wells is Cannot effectively
in handling no problem. produce deep wells
tubulars. to abandonment.
Easy to install System is depth Fairly flexible- Some difficulty in
downhole pressure limited because of convertible from analyzing properly
sensor for telemetering cable cost and continuous to intermittent without engineering
pressure to surface by inability to install to supervision.
cable. enough power plunger lift as well
downhole. declines.

26 | P a g e
 Availability of Not easily analyzable Easy to obtain Casing must
different sizes. unless good downhole pressure withstand lift
engineering know- and gradients. pressure.
how.
Lifting costs for high Gas and solids Sometimes Safety problem with
volumes generally production are serviceable with high pressure gas.
very low troublesome. wireline unit.
Crooked holes present Lack of production Crooked holes present
no problem. rate flexibility. no problem.
More downtime when Corrosion is not
problems are usually as adverse.
encountered
because of the entire
unit being downhole.
Casing size Applicable offshore.
limitations.
Cannot be set below
fluid entry without a
shroud to route fluid by
the motor. Shroud also
allows corrosion
inhibitor to protect
outside of motor.

Table 4: Design considerations and overall Comparisons [7]

Design considerations and overall Comparisons [7]

Consideration/System ESP Gas lift
Capital cost details Relatively low capital cost if electric Well gas lift equipment cost low but
power available. Costs increase as compression cost may be high.
horsepower increases. Central compression system
reduces overall cost per well.
Downhole Equipment Requires proper cable in addition to Good valve design and spacing
motor, pumps, seals, etc. Good design essential. Moderate cost for well
plus good operating practices essential. equipment (valves and mandrels).
Choice of wireline retrievable or
conventional valves.
Operating Efficiency Good for high-rate wells but Fair. Increases for wells that require
decreases significantly for <1000 small injection GLRs. Low for wells
bbl/day. Efficiency can vary from requiring high GLRs. Typically 20%,
40% in a low-rate well to 60% in a but range from 5 to 30%.
high-rate.
Flexibility of system Poor for fixed speed. Requires careful Excellent. Gas injection rate varied to
design. Variable speed drive provides change rates. Tubing needs to be sized
better flexibility. correctly.

Miscellaneous Requires a highly reliable electric power A highly reliable compressor with 95+%
problems system. System very sensitive to run time required. Gas must be properly
changes downhole or in fluid properties. dehydrated to avoid gas freezing.

27 | P a g e
Operating costs Varies. If high horsepower, high Well costs low. Compression cost
energy costs. High pulling costs result varies depending on fuel cost and
from short run life especially in compressor maintenance.
offshore operation. Repair costs often
high.
System reliability Varies. Excellent for ideal lift cases; Excellent if compression system
poor for problem areas (very sensitive properly designed and maintained.
to operating temperatures and electric
malfunctions).

Salvage value Fair. Some trade-in value. Poor Fair. Some market for good used
open market values. compressors and mandrels/valves.

System total Fairly simple to design but requires An adequate volume, high pressure,
good rate data. System not forgiving. dry, noncorrosive, and clean gas supply
Requires excellent operating practices. source is needed throughout the entire
Each well is an individual producer life. System approach needed. Low
with a common electric system. backpressure beneficial. Good data
needed for valve design and
spacing.

One of the factors to consider in artificial lift selection is the failure rates for the various systems
or the individual components of the systems. Figure 13 shows the run-life of ESP systems
versus their designed motor Horse Power (HP).

Figure 13; Run life of ESP systems. The figure is based on Centrilift data. [13]

28 | P a g e
It is important to note that data like this, which is based on a manufacturer’s experience, may
be subject to overestimation.

When planning a new well or field, one must be careful to compare too much with other run
life studies. The run life of a system is dependent on local conditions like scale potential,
temperature, sand production etc.

6.2.3.Selection by Net Present Value Comparison

A more thorough selection technique depends on the lifetime economics of the available
artificial lift methods. The economics, in turn, depend on the failure rates of the system
components, fuel costs, maintenance costs, inflation rates, anticipated revenue from produced
oil and gas, and other factors that may vary from system to system.

A typical NPV formula can look like this:

WI(QHC ×PHC −Cost−Tax)

NPV = ∑ni=1 (1+k)n
(Eq. 6.1)

Where:

WI = Work Interest (Talisman Energy Norge has 61% on Gyda)

Q = Oil rate

P = Oil price

Cost = All costs, operational (Opex) and capital (Capex)

Tax = Governmental taxes

k = depreciation rate of the project (percent)

To use the NPV comparison method, the user must have a good idea of the associated costs for
each system. This requires that the user evaluate each system carefully for the particular well
and be aware of the advantages and disadvantages of each method and any additional

29 | P a g e
equipment that may be required. Because energy costs are part of the NPV analysis, a design
for each feasible method must be determined before running the economic analysis to better
determine the efficiency of a particular installation [4].

7. ELECTRIC SUBMERSIBLE PUMPS (ESPS)

Electric Submersible Pump systems incorporate an electric motor and centrifugal pump unit
run on a production string and connected back to the surface control mechanism and
transformer via an electric power cable.

The downhole components are suspended from the production tubing above the well’s
perforations. In most cases the motor is located on the bottom of the work string. Above the
motor are the seal section, the intake or gas separator, and the pump. The power cable is
clamped to the tubing and plugs into the top of the motor.

As the fluid comes into the well it must flow past the motor and into the pump. This fluid flow
past the motor, aids in the cooling of the motor. The fluid then enters the intake and is taken
into the pump. Each stage (impeller/diffuser combination) adds pressure or head to the fluid at
a given rate. The fluid will build up enough pressure, as it reaches the top of the pump, to be
lifted to surface and into the separator or flow line.

30 | P a g e
Figure 14; Basic ESP

31 | P a g e
 Figure 15; ESP surface system.

Figure 15 shows an example of a full ESP system. Since this study is more about pump
performance and sizing, the surface equipment will not be described in detail. However, the
importance and complexity of this must not be forgotten in an ESP design. The surface
controller provides power to the ESP motor and protects the downhole ESP components. Motor
controller designs vary in complexity from the very simple and basic to the very sophisticated,
which offers numerous options to enhance the methods of control, protection and monitoring
of the operation.

32 | P a g e
Submersible systems have a wide performance range and are one of the more versatile lift
methods. Standard surface electric drives outputs from 150 to 150,000 bbl/d (24 to 24,600
m3/d) and variable speed drives add pump flexibility. High GOR fluids can be handled, but
large gas volumes can lock up and destroy pumps.

Corrosive fluids are handled by using special materials and coatings. Modified equipment and
procedures allow sand and abrasive particles to be pumped without adverse effects. [15]

7.1. Centrifugal pump

The ESP is a multistage centrifugal pump. A cross section of a typical design is shown in figure
16.

Figure 16; the inside of a centrifugal pump.

The shaft is connected to the seal-chamber section and motor. It transmits the rotary motion
from the motor to the impellers of the pump stage. The shaft and impellers are keyed, and the
key transmits the torque load to the impeller.

The stages of the pump are the components that impart a pressure rise to the fluid. A stage is
made up of a rotating impeller and a stationary diffuser.

33 | P a g e
 Figure 17; Shaft with the rotating impellers attached

The stages are stacked in series to incrementally increase the pressure to that calculated for the
desired flow rate. Figure 17 shows the flow path. The fluid flows into the impeller eye area and
energy, in form of velocity, is imparted to it as it is centrifuged radially outward impeller
passageway. Once it exits the impeller, the fluid makes a turn and enters the diffuser
passageway. As it passes through this passageway, the fluid is diffused, or the velocity is
converted to a pressure. It then repeats the process upon entering the next impeller and diffuser
set. This process continues until the fluid passes through all stages, and the design discharge
pressure is reached. This pressure increase is often referred to as the total developed head
(TDH) of the pump.

There are two styles of stages for the range of flow rates in which ESPs operate. A radial stage
has the flow entering the impeller or diffuser parallel to the axis of the shaft and exiting
perpendicular to the shaft. This is often referred to as a “pancake” or “mushroom” stage because
of its flat shape. The second style is the mixed flow stage which has the flow exiting the
impeller at an angle less than 90° to the shaft (see stage in fig 18).

The mixed flow design handles larger flow rates than the radial and is not that vulnerable to
free gas and particles [4].

34 | P a g e
A Key feature for both styles of stages is the method by which they carry their produced axial
thrust. Usually, the pumps that are under a 6 inch diameter are built as “floater” stages. On
these, the impellers are allowed to move axially on the pump shaft between the diffusers. They
typically run in a down-thrust position and at high flow rates, they may move into up-thrust.

To maintain the optimum flow path alignment between the impeller and its diffuser, the
impeller is designed to maintain a down-thrust position through its operating range (figure 18).

Figure 18; ESP operating range

The manufacturers give the pump performance characteristics on the basis of 1 stage, 1,0 SG
water at 60- or 50-Hz power. A typical performance graph is shown in figure 19. The head,
brake horsepower (BHP), and efficiency of the stage are plotted against flow rate on the x-axis.
Pump efficiency is given by:

[Q×TDH×SG]
ηp = (C×BHP)

Where:

Q = flow rate

35 | P a g e
TDH = Total Head Developed

SG = specific gravity

BHP = Break horsepower

C = constant = 6, 75 (when Q = m3/d and TDH = m)

The head/flow curve shows the head or lift, measured in feet or meters, which can be produced
by one stage. Because head is independent of the fluid SG, the pump produces the same head
on all fluids, except those that are viscous or have free gas entrained. If the lift is presented in
terms of pressure, there will be a specific curve for each fluid, dependent upon its SG.

The highlighted area on the graph is the manufacturer’s recommended operating range. It
shows the range in which the pump can be reliably operated. The left edge of the area is the
minimum operating point, and the right edge is maximum operating point. The best efficiency
point (BEP) is between these two points, and it is where the efficiency curve peaks. The shape
of the head/flow curve and the thrust characteristics curve of that particular stage determines
the minimum and maximum points. The minimum point is usually located where the head
curve is still rising, prior to its flattening or dropping of and at an acceptable down-thrust value.
The location of the maximum point is based on maintaining the impeller at a performance
balance based on consideration of the thrust value, head produced and acceptable efficiency.

Figure 19; Standard pump curves for head, efficiency and BHP [19].

36 | P a g e
 7.2. Seal Chamber section
The component located below the lowest pump section and directly above the motor, in a
standard ESP configuration, is the seal chamber section. It is basically a set of protection
chambers connected in series or in some special cases in parallel. This component has several
functions that are critical to the operation and run life of the ESP system, and the motor in
particular.

 It protects the motor oil from being contamination by the wellbore fluid.
 It allows for pressure equalization between the interior of the motor and the wellbore.
 It also absorbs the axial thrust produced by the pump and dissipates the heat that the
thrust bearing generates.

Figure 20; Seal.

Figure 20 shows a mechanical seal which is generally located at the top of each protection
chamber and is used to prevent well fluid from migrating down the drive shaft.

7.3. The Motor

The ESP motor is a two-pole, three-phase, squirrel cage, induction design. A two- pole design
means that it runs at 3600 rpm at 60 Hz power or roughly 3500 rpm actual operating speed. It

37 | P a g e
operates on three-phase power at voltages as low as 230 and as high as 5000. Generally, the
length and diameter determine the motors HP rating. Because the motor does not have a power
cable running along its length, it can be manufactured in diameters slightly larger than the
pumps and seal chamber sections and still fit in the same casing bores.

7.4. The power cable

The ESP power cable transmits the required surface power to the ESP motor. It is a specially
constructed three-phase power cable designed specifically for downhole well environments.
The cable design must be small in diameter, protect from mechanical abuse, and impervious to
physical and electrical deterioration because of aggressive well environments. They can be
manufactured in either round or flat configurations (figure 21 and 22). The round design is the
best conductor, but the flat design is often used beneath the ESP packer and along the pump
and seal section because of the small space between ESP and casing.

1=Armour, 2=Jacket, 3=Basic insulation, 4=Physical filler, 5=Conductor

Figure 21; round design

1=Armour, 2=Braid, 3=Barrier layer, 4=Jacket, 5=Conductor/insulation gas block,

6=Conductor

Figure 22; flat design

38 | P a g e
 7.5. ESP run life
ESP run lives are dependent on numerous variables broadly characterised as; equipment,
operation and operating environment. A combination of these factors can produce significant
variation in ESP survival times, as presented in Figure 23.

Figure 23; Factors acting on ESP run life [13]

The reliability model for ESPs is described as the “bathtub” concept and uses three stages in
the life of an ESP:

 Stage one Infant mortality ESP fails to start at installation.

 Stage two In-service failures Operational issues.
 Stage three Wear out Failure due to pump wear out.

The failure rates experienced in each of the three stages are not related and must be analysed
separately.

Stage one failures occur within 2 days or less of operation. These are therefore represented as
a percentage of installations that are predicted to not start. A typical cause would be damage
on installation usually when running in hole (RIH), incorrect electrical connection or foreign
objects left in the well damaging the stages of the pump.

39 | P a g e
Stage two failures are time independent; this is the field operation of the equipment. Electrical
failures are common due to insufficient cooling of a unit which occurs at very low flow rates,
through gas locking and dead heading of the pump-all of these are avoidable under careful
supervision. A particularly high-risk time is at start up when the shaft can snap and undesirable
operating ranges may need to be traversed. Pressure cycling can also cause the cabling to the
pump to fail.

Stage three failures are analysed less frequently with strip downs as pumps generally are
expected to fail at this point or have been replaced as part of a proactive workover plan. Many
of the components have a limited life, especially the seals, which will degrade over time [12].

The factors on which the run life is dependent on are listed below:

Design and Sizing

Proper sizing of the ESP unit is the first factor in achieving a long run life. The unit must be
sized to operate within the recommended flow range. Well productivity data must be accurate
to properly size the equipment. The consequences of improper sizing are that the ESP will be
running outside of operating range causing accelerated pump wear, the risk of motor burn out
from excessive gas locking or very low flow rates. Inaccurate fluid data can lead to a pump
sized for the wrong conditions.

Operating Practice

Poor operating practices are a major cause of failure for ESPs. These can be as a result of lack
of knowledge in operating the units or an unexpected change of the operating environment.

Downhole information can be used to provide a better perspective of ESP operation and
performance. Only 2% of ESPs in the world have downhole sensors and even those with the
data often neglect to use it to control the pumps. Detailed real time information concerning the pump
pressures and temperatures the system is experiencing downhole can be used to help protect, control
and optimize the operation of the ESP.

BHT Temperature

Bottomhole temperatures greater than 105 °C is considered a high temperature application for
ESPs. The motor assembly will need to be checked for clearance at the higher temperatures.

40 | P a g e
The consequence of not taking these measures will be a shorter component life or reduced
MTTF (Mean Time to Failure).

Free Gas

As ESPs are designed to pump liquids and not gas, breakout of free gas or alternating slugs of
liquid and gas can lead to operational difficulties. As fluid velocity decreases past the motor,
cooling will become less efficient, and the danger of the motor overheating and burning out
increases. In the extreme, as the proportion of free gas increases, the pump begins to lose head
and spin empty of fluid in a condition called gas locking.

Viscosity

High fluid viscosity can cause many problems. As the specific gravity of the fluid increases,
so does the pump break horsepower requirement. High viscosity also reduces the pump’s
ability to lift the fluid and its efficiency, as the viscous fluid produces more frictional pressure
loss in the tubing causing the pump to work much harder. The viscosity of produced fluids may
change with the application of shear by the pump; this may alter over a range of water cuts.
Tight emulsions can be formed under certain conditions.

Corrosion

Corrosion from CO2 and H2S can affect the ESP unit by eroding electrical connections, seals
and fastenings long before impeller performance is degraded. Appropriate material selection
can avoid these issues.

Sand Abrasion

Sand production can be detrimental to ESP performance by reducing pump efficiency through
abrasive wear to the stages. More immediate failure is due to increased pump shaft vibration,
which in turn leads to mechanical failure of the seals and motor burn out due to the subsequent
fluid migration.

41 | P a g e
The most effective strategy is to eliminate or reduce sand production. Sand production can be
managed through controlled start up and an understanding of sand mobilisation rates. The same
sand can be produced repeatedly through the pump without making it surface.

Damage to impellers and stages can be reduced by appropriate material selection and an
abrasion resistant pump design which provides support and radial shaft stabilization.

Foreign Material

The production of foreign material can cause damage or failure of an ESP. Although rare, a
foreign object can jam the pump resulting in motor burn out, more commonly the material will
damage the impellers thus reducing the lift efficiency of the pump.

Deposition

Scale, asphaltenes, paraffin and hydrates can all deposit in ESPs. The result can be blocked or
limited pump inflow, reduced efficiency of pump stages or locking of stages, with the
consequences being reduced efficiency and the associated danger of motor burn out.

Electrical Failure

Electrical failure can happen at surface or downhole. Problems at surface such as overload of
the controller or transformer are easier to rectify than those downhole that interrupt the power
source and require a workover intervention to change out ESP.

Old Age

Even if the ESP has been operated within the design envelope and care has been taken
operationally, the time will come where certain components reach failure point. The hardware,
stages and bearings are usually over designed so the failure is most likely to come from
‘consumable’ items. Seals will degrade over time, motor oil deteriorates, o-rings and
connections all have shelf lives and electrical components within the pump and downhole
monitoring package will fail. However, there are many examples of ESPs that have exceeded
run life targets of over 5 years in operation.

42 | P a g e
Reliability Issues Specific to High Horse Power Units

Higher HP units are exposed to greater risk. A higher HP unit contains more motor sections
and is therefore physically longer than other units. Installation can lead to the mechanical
damage of units which puts them into the infant mortality category of the reliability model. The
longer length of the unit, the higher the risk of damage.

Dogleg severity and deviation limits will be required to be more stringent than for the shorter
models. Increased physical protection can be supplied by running pumps in an enclosed pod to
provide defence against mechanical damage whilst RIH. This system is much easier to
workover and carries much lower operational risks.

High horsepower pumps are made up of several lower horsepower pumps run in series and are
therefore dependant on one another. This dependency intrinsically reduces the reliability of the
whole system. Reliability is also reduced by the requirement for higher power and torque to be
supplied to one motor which then feeds the others [12].

7.5.1.Case studies
Beatrice (Talisman UK) [10]
The graph below, Figure 24, represents the frequency of the run time of pumps installed in the
Beatrice Field, where Talisman UK is the Operator. This data shows that 50% of the installed
pumps (P50) run for up to 470 days without failure. The available data was for ESPs with a HP
range between 165 and 685 HP. Most of the Beatrice ESPs are within two HP intervals (201-
300) and (401-500). In order to obtain a more accurate analysis, data represented by these two
ranges was used to determine the P50 run life instead.

43 | P a g e
 Figure 24; ESP run life on Beatrice. The figure shows data from two HP intervals [13].

This Beatrice data analysis covers 54 operating units which is a respectable sample for an
individual field. On first inspection, it would seem that the reliability of the ESPs increases
with HP, which is contrary to the model that suggests increasing motor complexity is related
to a decrease in ESP reliability. However, the range of motor sizes is relatively narrow; most
units in the higher HP range are less than 450 HP and the lower range being mainly 250 HP
units. Attributing difference in the operating run life over this small range of HPs does not seem
reasonable. The run life duration is being affected by some other factors, namely the operation
and operating environment of the pumps.

Wytch Farm
The table below, Table 5, represents the frequency of the run time of pumps installed in the
Wytch Farm Field, where BP is the operator. This data shows that 50% of the installed pumps
(P50) run for up to 1000 days without failure. The available data was for ESPs of a motor size
range between 840 HP and 1170 HP. The Wytch Farm data set of 27 units is smaller than the
Beatrice data set, but covers a much wider range of motor sizes. The trend here supports the
theory that as HP increases, reliability decreases.

44 | P a g e
 Table 5; Wytch Farm ESP run life [13]

ESP HP 840 900 1170

P50 Run Life, days 1216 902 859
Average Run Life, days 1073 1108 847

The Wytch Farm run lives are substantially longer than those experienced on the Beatrice field
even though much larger horse powered motors are employed.

In order to select which of the two run life results would be more suitable to represent the
installation conditions in the Gyda field, a comparison between the properties of both fields
against Gyda has to be done. The biggest issue compared to Wytch Farm and Beatrice is the
temperature. The two fields have reservoir temperatures of 70 and 80 °C, while Gyda has a
reservoir temperature of 160 °C. Scale potential and possible sand production as a result of the
high-pressure drawdown imposed by the pumps has been identified as two main risks.

Centrilift, which is the ESP supplier, has given an estimated lifetime of approximately two
years with a dual ESP solution (described in next section). It must be kept in mind that a
manufacturer’s estimate may be too optimistic, and it is important that TENAS performs or
receive a proper analysis of the ESPs running in Gyda conditions before making a decision.

7.6. ESP design for Gyda

Because of the limited lifetime of ESPs, the suggested design for Gyda is a dual system. This
means that two separate systems are installed, one upper and one lower. Only one ESP system
operates at the time. The other unit is held as backup until it either fails or is shut down
voluntarily.

Dual ESP lift systems enable cost-effective production in applications where rig availability
may be at a premium and where the cost of workover impacts the overall profitability of the
well.

45 | P a g e
 Figure 25; Dual ESP system design for Gyda

The ESPs are planned to be built into shrouds. This is done to seal off the entire system from
the casing. The casings in A-19 and A-26 where not originally designed to handle production
and interaction with produced fluids. When increased pressure variation from the pumps also
was added, it was decided that the sealed system was superior.

To divert the flow between the upper and lower system, Automatic Diverter Valves (ADV) are
placed above each pump section. The valves will be in a closed position when there is a pressure
build-up from beneath, and will open when there is a pressure build up from above. This means
that when the ESP is operating the valve will stay in a closed position and opens when the
pump is turned off. Therefore, the production flow can now be directed around or through each
ESP depending on which one is operating. The ADV also protects the pump against solids and

46 | P a g e
fluid fall-back when it is turned off, which again increases the run life of the pump. The design
allows for bullheading, for killing the well or scale squeezing. A schematic of the ADV
operating is shown in figure 26 [22].

Figure 26; ADV operating schematic [22]

7.7. ESP secondary effect

Gyda is divided into 3 main units A-, B- and C-sand (chapter Gyda field), where the upper A-
sand has the highest permeability.

When installing ESPs, it is expected that the pumps will draw their near wellbore area pressure
down with up to 200 bar. An effect of this will be that oil from the low permeability areas will
migrate into the high permeability areas. Figure 27 shows how oil from the B-sand and lower
A-sand migrates into the high permeability upper A-sand. There are large oil reserves in the
low permeability zones that initially never would be produced, and this secondary effect can
increase oil recovery significantly.

47 | P a g e
 Figure 27; ESP secondary effect. Pressure drawdown makes oil migrate into the high perm.
Zone

This effect is not included in the production forecast later in this study, because it has not been
accounted for in the initial reservoir model. However, it can be kept in mind that the result from
the production forecast will be conservative and that there is a great possibility of a higher
production because of this secondary effect.

8. GAS LIFT
Gas lift is the process of injecting gas in the annulus between tubing and casing where it will
enter the tubing via a gas-lift valve located in a side pocket. The gas will then reduce the weight
of the produced fluid column, which will lower the bottomhole pressure. Reservoir fluid will
then experience lower resistance to flow, resulting in increased flow rates and increased
production.

48 | P a g e
 Figure 28; Gas lift. Gas is pumped down the annulus and into the tubing [18].

Gas lift is the artificial lift method that most closely resembles the natural flow process. The
only major requirements are a supply of pressurized injection gas. Normally, the lift gas is
supplied from other producing wells, separated from the oil, run through a gas compressor and
pumped in the annulus at high pressure. The gas from the producing well is then recovered
again, recompressed and re-injected.

However, the gas compressing process is power consuming and expensive [16].

Figure 28 shows a typical continuous flow gas lift installation. Other methods are: Intermittent
gas lift:

49 | P a g e
If the gas is injected in intervals it is called intermittent gas lift. This can be used in low
producing wells where one wants to enable some liquid to build up in the bottom of the well
before one injects a slug of gas which carries the fluid to surface [18].

Dual gas lift:

Because of limited drilling slots on rigs, an alternative to drilling two separate wells to drain
different reservoir sections is to have two independent tubing completions in the same well. If
these require gas lift, the gas can be supplied from a common casing and injected into the
different gas lift valves (GLV). Another way is to inject the gas in one string and produce from
the other. This type of completion is called a dual gas lift.

Self-lift:

One alternative to the process of injecting the gas from surface is to simply let it flow naturally
from the gas reservoir located above the oil zone. This type of gas lift is called self-lift, auto
lift or natural lift. The well must penetrate both the gas cap and the oil zone to achieve this. A
gas cap or reservoir capable of sustaining the well with sufficient rates for the lifetime of the
well is needed for this completion. Self-lift completions eliminate the need for surface gas
compressors and other gas lift equipment. Such wells can be fitted with downhole flow-control
valves and permanent monitoring equipment which qualifies the wells as intelligent wells [16].

8.1. The Unloading Process

After a well is completed or worked over, the fluid level in casing and tubing is usually at or
near the surface. The gas lift pressure available to unload the well is generally not sufficient to
unload fluid to the desired depth of gas injection. This is because the pressure caused by the
static column of fluid in the well at the desired depth of injection is greater than the available
gas pressure at injection depth. In this case a series of unloading valves is installed in the well.
These valves are designed to use the available gas injection pressure to unload the well until
the desired depth of injection is achieved. Figure 29 shows the unloading process in a
continuous gas lift system.

50 | P a g e
 Figure 29; the unloading process.

The figure shows a well with 3 gas lift valves, the two uppermost are called unloading valves,
while the lowest one is called operating valve. When gas reaches the first unloading valve, gas
is injected into the tubing, as shown in part B of Fig 29. The liquid in the tubing get aerated
and the static tubing pressure at the valve depth decreases to a stabilized low value that
corresponds to the gas liquid ratio (GLR).

The lower valves are still open and the liquid level in the annulus continues to drop.

When the liquid level in the annulus reaches the next unloading valve, gas will be injected
through the valve. This is the most critical moment in the unloading process, because both
unloading valves inject gas at the same time as shown in Part C. The upper valve has to be
closed in order to move the injection point down to the operating valve and ensure that gas is
injected at a single point only. Proper design and setting of the unloading valves ensure that
the shallower valve closes just after the next lower valve starts injecting gas, as shown by part
D.

As the middle valve continues injecting gas, the tubing pressure at that depth falls and the
annulus fluid level continues to drop. If the unloading string is properly designed, the stable
liquid level in the annulus will be just below the lowest valve, which is the operating valve.
When gas reaches the operating valve, gas will be injected into the tubing. Then it is very
important that the middle valve closes, as shown part F. By now the objective of the unloading
process has been met and gas is injected through the operating valve only [2].

51 | P a g e
 8.2. Gas lift performance curve
The gas lift performance curve is a plot of the well’s liquid rate vs. the gas injection rate for a
given surface gas injection pressure and shows the producing system’s response to continuous
flow gas lifting (Fig. 30). The figure indicates that at low injection rate, any increase in the gas
volume increases the well’s liquid output. As injection rates increase, the rate of liquid volume
increase falls off and the maximum possible liquid rate is reached. After this maximum any
additional gas injection decreases the liquid production. In this region of high gas injection
rates, multiphase flow in the tubing is dominated by frictional effects. Consequently, bottom-
hole pressure starts to increase and liquid inflow to the well diminishes.

Figure 30; Gas lift performance curve. The figure shows the total production rate plotted
against the rate of lift gas injected [14].

8.3. Gas lift valves

Unloading valves
The Injection Pressure Operated (IPO) valve is the most commonly used type of unloading gas
lift valve, and is currently used on Gyda. Other types worth mentioning are Production Pressure
Operated (PPO) valves, balanced bellows valves, balanced flexible sleeve valves, and Pilot
valves.

52 | P a g e
A standard IPO valve contains a nitrogen pre-charge chamber and a flexible bellows assembly,
which provide the closing force of the valve. The axial position of the stem determines if the
valve is open or closed for gas injection. When injection gas pressure exceeds the closing force
the bellows compresses, thus lifting the valve stem off its seat, allowing gas to be injected
through the valve. The schematic of a conventional IPO unloading valve is shown in Fig. 31.
The reverse flow check valve consists of a spring-loaded dart and a seat. When the check valve
is closed tubing- pressure from below push the dart against the seat. Despite the check valve,
tubing pressure can act on the valve stem due to pressure trapped between the port and check
valve and unavoidable imperfection in the check valve seal.

Figure 31; conventional unloading valve [24].

By writing the force balance for the valve stem, the conditions for opening and closing can be
found. In the closed position nitrogen dome pressure P d acts on the area of the bellows Ad and
provides enough force to keep the stem against the port. All other forces try to open the valve,
the largest force comes from the injection pressure Pi, which acts on the doughnut shaped area
of the bellows Ad minus the area of the port Ap. A much smaller force comes from the
production pressure Pp that acts on the port area Ap of the stem tip. The valve opens when the
sum of the opening forces exceeds the closing force:

53 | P a g e
Pi (Ad − Ap ) + Pp Ap > Pd Ad (Eq. 8.1)

Solving this equation for injection pressure Pi when forces are in equilibrium:

P A P A
Pi = (A d−Ad ) − (A p−Ap (Eq. 8.2)
d p d p)

To simplify, introducing the term R = Ap/Ad, gives the equation for the injection pressure
necessary at valve depth to open the valve:

P d pP R
Pi = (1−R) − (1−R) (Eq. 8.3)

As seen from the equation (8.3), to open the valve from the closed position depends not only
on injection pressure, but also on production pressure. If production pressure is constant, the
valve will open when injection pressure exceeds the value calculated by this equation. When
the valve starts to open injection pressure will act on the whole area of the bellows and
completely lift the stem off the seat. When the valve is fully open a new force balance can be
written. Closing force is as before from the dome pressure P d, acting on the bellows area Ad.
The opening force is provided by only injection pressure P i acting on the bellows area Ad. The
closing equation is therefore:

Pi Ad = Pd Ad (Eq. 8.4)

Pi = Pd (Eq. 8.5)

The valve will close when injection pressure becomes lower than the dome pressure. A
graphical presentation of the operating principle is shown in Fig. 32. When production pressure
is higher than injection pressure, the valve is closed by the check valve, that is shown by the
area above the line Pp = Pi. At injection pressures lower than the dome pressure, the valve will
be closed, as shown by the shaded area. Opening of the valve occurs along the bold line where
equation (8.3) is in equilibrium. Closing occurs where dome pressure equals injection pressure,
as shown by the vertical line and equation (8.5). In the triangular area between the opening line
and the closing line, the state of the valve depends on its previous state. This design allows for

54 | P a g e
flexibility should production conditions change, such as water-cut increase. The disadvantage
of this type of valve is that the maximum depth of injection is reduced for each unloading valve
used, because casing pressure has to be reduced in order to close an unloading valve [2].

Figure 32; IPO gas lift valve operating principle [2]

Operating valve
The lowest valve is often referred to as the operating valve or orifice valve. For a continuous
gas lift system, it is recommended to use a "Nozzle-Venturi valve" which will provide a more
constant gas rate than the old “square-edged orifice”.

In a conventional square edge orifice, the gas rate through the valve increases as differential
pressure over the valve increases gradually until critical flow is achieved at a critical pressure
ratio of about 0.55. At critical flow, supersonic velocity is reached in the orifice, and a further
increase in differential pressure will not cause increased gas rate. A typical continuous gas lift
installation operates in the subcritical range; this implies that that gas injection rate will change
due to inevitable pressure fluctuations at valve depth. When tubing pressure at valve depth
increases, the injected gas rate will decrease and vice versa, since the valve is in the sub critical
range. This behavior is opposite to the basic principle of steady continuous gas lifting, which
is that there should be more gas injected when it is needed.

Therefore, square edge orifice valves are usually not recommended due to instability problems
and occurrence of heading in tubing and casing. The orifice size or cross-sectional area of the
throat will have to be sized according to the gas rate to be injected to achieve critical flow. A
common mistake in gas lift design is to install a too large orifice size which would cause a
lower percentage of critical flow. To prevent the problems associated with sub critical flow and

55 | P a g e
instability, replacing the square edge orifice with a converging-diverging venturi nozzle as
shown in Fig 33, will ensure that critical flow is achieved at a significantly lower differential
pressure ratio. A venturi valve will achieve critical flow at as little as roughly 0.9 in differential
pressure ratio as shown in Fig 34. This means that with a venturi valve the gas injection rate
will stay stable and independent of tubing pressure fluctuations, and prevent heading and
instability. The disadvantage of this valve is that it is less flexible for future changes, since the
injection rate through it is almost fixed, and that could possibly lead to a lower production rate
if a higher production rate is achievable at another rate.

For preventing backflow into the tubing/casing annulus, they are equipped with standard
reverse flow check valves.

Figure 33; cross section of square edge orifice and venturi valve [2].

Figure 34; Gas passage characteristics comparison [2].

56 | P a g e
 8.4. Gas lift completion procedure
The tubing is fitted with a side pocket mandrel, where the side pocket can have a gas- lift valve,
a chemical-injection valve or similar. The gas-lift valve can either be preinstalled or it can be
placed in the side pocket by means of wireline. If it is not preinstalled the side pocket will
contain a dummy valve which isolates the tubing from the annulus. Setting and pulling a gas-
lift valve is normally done by slickline. But if the well has a deviation of over 65 degrees, an
electrical wireline in combination with a well tractor is used. In both cases, a kickover-tool
(KOT) is run in the well.

The KOT is lowered into the hole and a location finger will locate an orientation groove in the
side pocket mandrel. It is important to be careful in the transition between pup joint and side
pocket as the location finger can get stuck and one can risk trigging the KOT prematurely.
Pulling on the wireline with a predetermined force will orient the KOT in the right direction.

Pulling further will shear a pin in the tool allowing the arm with the valve to kick over. It can
then be lowered into the empty side pocket if the well is vertical or pushed by a stroker tool if
it is deviated. One should then pressure up the tubing or bleed down the annulus to make sure
that the valve is properly latched into the side pocket. Hopefully, the pressure difference will
be sufficient to push the valve in place.

57 | P a g e
 Figure 35; Valve installation with KOT. [24]

When the valve is latched into the side pocket one can pull on the wireline with a predetermined
force or activate the stroker and another pin will shear, freeing the running tool from the gas-
lift valve. Further pulling on wireline or activation of stroker will shear another pin in the KOT.
The tool string can then be pulled out of the well. Figure 35 shows the running process. The
sequence of pulling a valve is identical, with a pulling tool replacing the running tool [24].

9. PROSPER
PROSPER is the industry standard single well performance design and optimization software,
it can model most types of well completions and artificial lifting methods. PROSPER is used
by major operators worldwide. The software allow building of well models with the ability to
address all variables such as well configuration, fluid characteristics (PVT), multiphase VLP

58 | P a g e
correlations and various IPR models. Tuning of the models is possible by matching real field
production data, the benefit of matching is the ability to model different scenarios with
increased accuracy.

Figure 36; PROSPER front display.

Calculation of VLP using multiphase flow correlations with evaluation of VLP variables is the
major application of the PROSPER software. Sensitivity analysis on future changes of
parameters that affect VLP and IPR are easily assessed.

The artificial lifting methods available are:

 Electric Submersible Pump - ESP

 Gas Lift
 Hydraulic Pumps - HSP

59 | P a g e
  Progressive Cavity Pumps - PCP
 Jet Pumps
 Sucker Rod Pumps
 Multiphase Pumps
 Injection of diluents
 Gas lift with coiled tubing

A full range of well types can be modelled in PROSPER including gas, oil, water, condensate
and steam. Different configurations such as angled, multi-layer and multi-laterals can also be
modelled. A full range of IPR models can be used in Prosper including: PI entry, Vogel,
Composite, Fetkovich, Jones, horizontal well model plus several others. Various completion
configurations such as gravel pack, open, cased and perforated hole are also available. [21]

9.1. Basic theory of PROSPER

To predict pressure and temperature changes from the reservoir, along the wellbore and flow
line tubular it is necessary to accurately predict fluid properties as a function of pressure and
temperature.

The Black Oil PVT model is used for the vast majority of applications.

Different applications can be chosen for The Black Oil model in PROSPER. It can be tuned to
a retrograde condensate model, dry and wet gas model, or an oil and water model. The oil and
water model which is used for this study, takes the surface production of oil and associated gas
together with the water cut to determine the well mass flow rate. PVT correlations are used to
find the amount of gas at each pressure and temperature. Bo, Bg and Bw are evaluated at each
calculation step to find the phase densities.

When both basic fluid data and some PVT laboratory measurements are available, the program
can modify the Black Oil model correlations to best-fit the measured data using a non-linear
regression technique. When detailed PVT laboratory data is provided, it can be entered in table
format.

The second option is the Compositional model. This is used when a full Equation of State
(EOS) description of the fluid is available and all the PVT can be obtained from a Peng-
Robinson or a Soave Redlich Kwong description of the fluid phase behavior. The EOS method

60 | P a g e
relies on empirical correlations for predicting density, viscosity etc. and should only be used
for specific specialized applications.

VLP correlations describe various methods of calculating pressure losses in inclined pipes.
While the basic form of the pressure loss equation is the same for all correlations, the treatment
of multiphase friction and gas/liquid slip (holdup) varies considerably. The various VLP
correlations were designed around specific sets of lab data. They can therefore be expected to
perform best for field conditions that approximate the experimental conditions.

There is no universal rule for selecting the best correlation for a given well. The correlation
must be selected on the basis of flow regimes and closeness of fit to measured pressure data.
In making a selection the purpose of analysis should be considered. A slug flow correlation
that closely matches current production may not be applicable if, for example, the GOR is
expected to increase greatly in the future.

The IPR describes pressure drawdown as a function of production rate. The drawdown is a
complex function of pressure drawdown, fluid PVT properties, formation permeability
(absolute and relative), effective overburden etc. For practical engineering purpose, a number
of IPR models have been developed and implemented within PROSPER. For this study the PI
Entry model is used [6].

9.2. Building a base model for A-19 and A-26

The first step when modelling a new well in PROSPER is to fill out a system summary like
figure 37 shows. The Black Oil model, with the oil and water option describing the fluid, is
used for both A-19 and A-26. It is also here the choice of artificial lift method is made, which
allows us to fill in information and design a system at a later stage.

61 | P a g e
 Figure 37; System summary.

9.2.1.PVT
The next step is to fill in the PVT data. First the basic data for the Black Oil model for the well
is entered (table 6). Second, PVT test data is entered and matched to the Black Oil correlations.

The correlations are mathematical best fits to sets of lab measurements. They represent the
PVT behavior of average hydrocarbons. Each individual fluid sample will behave similarly on
average, but not exactly as predicted by the correlation. By comparing the values predicted by
the correlation and measure lab data, adjustment factors for the correlation can be found that
minimize the overall difference.

PROSPER uses a non-linear regression to do this by applying a multiplier (Parameter 1) and a

shift (Parameter 2) to each correlation. Figure 38 shows how the software matches the data.
Parameter 1 should be close to one and Parameter 2 close to zero. Standard deviation represents
the overall closeness to fit and should be as small as possible.

62 | P a g e
 Figure 38; PVT correlations.

Table 6; General PVT input for A-19 and A-26

Input parameters A-19 and A-26

Solution GOR, Sm3/Sm3 159,4
Oil Gravity, kg/m3 804,2
Gas Gravity, kg/m3 1,07864
Water salinity, ppm 172995
Mole percent H2S 0
Mole percent CO2 1,77
Mole percent N2 0,868

9.2.2.IPR
When the PVT data has been properly matched, one can start making the IPR curve. As
mentioned in the well performance chapter there are a lot of models for making the IPR. The
most used model for Gyda wells is the simple “PI entry”. This model is based on equation 4.4

63 | P a g e
which generates a straight line above the bubble point. The Vogel empirical solution (Eq. 4.5)
is used below bubble point. The Productivity index entered is used to calculate the IPR. The
IPR rates are always given as liquid rates. Hence the PI refers to liquid rate.

The data in table 7 and 8, which is entered into the IPR section, is based on an Eclipse reservoir
simulation made by the TENAS reservoir department. The reservoir simulation covers a “base
case”, where no artificial lift method is installed, and an “ESP case”. There has not been made
a prediction for a “gas lift case” yet, but it is assumed that a gas lift installation would lower
the static BHP with 20 bars compared to the “base case”. “Base case” water cut is also assumed
for the gas lift.

The water cut is lower for the “ESP case”. ESPs will drastically draw down the pressure in near
wellbore area. This makes the oil to expand, thus less water is produced.

The GOR is approximately the same for both the simulated scenarios, so the same GOR is
assumed also for gas lift. Because A-26 is currently dead, there is no “base case” simulation
for production and hence no prediction of GOR or water cut.

However, since it is assumed that the GOR is independent of lift method, the same GOR as for
ESP production is assumed. The “base case” water cut is based on the latest well test.

All the base simulations (PROSPER) in this study are based on the predicted reservoir
conditions for May 2010 which is the planned finish of artificial lift installation in A-19 and
A-26.

Table 7; IPR input for A-19

IPR data for A-19, predictions for May 2010

Reservoir pressure, barg 400
Reservoir temperature, °C 155
Water cut, % 53, 44 for ESP
Total GOR, Sm3/Sm3 140
PI, Sm3/d/bar 6

64 | P a g e
 Table 8; IPR input for A-26

IPR data for A-26, predictions for May 2010

Reservoir pressure, barg 400
Reservoir temperature, °C 155
Water cut, % 86, 81 for ESP
Total GOR, Sm3/Sm3 160
PI, Sm3/d/bar 9

9.2.3.Equipment
For PROSPER to be able to calculate the pressure and temperature profile along the well,
completion, survey and temperature data is needed.

For A-19 and A-26 the following survey data is entered:

Figure 39; survey data for A-19. The well path is plotted to the right.

65 | P a g e
 Figure 40; survey data for A-26. The well path is plotted to the right.

For implementation of gas lift or ESP in the two wells a full workover is necessary. ESPs are
a part of the tubing, and the wells does not contain side pockets so that gas lift valves can be
installed by wireline. However, the geometry of the production tubing would look very much
the same as for the old wells. Therefore, the current well completion is used for the PROSPER
simulations. Inner and outer diameter of casings, tubings and liners are put into the model.
Inner diameter of restrictions like the downhole safety valve is also accounted for. A tubing
and casing inside roughness of 0,0006 inch is used. Figure 41 and 42 shows the current
completion schematics of the two wells.

66 | P a g e
Figure 41; Completion schematic of A-19

67 | P a g e
 Figure 42; Completion schematic of A-26

For both wells the temperature profile in table 9 is used. Default values for heat capacities are
entered.

68 | P a g e
 Table 9; Temperature profile

Position Temperature
Wellhead 10 °C
Seabed 6 °C
Top perforation 155 °C

9.2.4.Results
Given these input parameters PROSPER gives a production profile, with no artificial lift, for
A-19 and A-26 in May 2010. A-19 has an oil production of 2381 bbl/d and A-26 produces 234
bbl/d of oil. The reason why we see a production in A-26 could be that there is enough pressure
support from injectors and in 2010 A-26 actually can produce. However, the well must be
considered as dead because it doesn’t take more than a small change, which is well within the
margins of error, in water cut or reservoir pressure for there to be no production point again
(figure 45). Figures 43 and 44 show the production point in May 2010, which is the intersection
between the VLP and IPR curve.

Figure 43; production point of A-19 in May 2010

69 | P a g e
 Figure 44; production point of A-26 in May 2010

Figure 45; A-26 with no intersection between IPR and VLP (situation today).

70 | P a g e
 10. GAS LIFT DESIGN
When performing a gas lift design, the best compromise between a number of objectives are
sought:

 Find optimum production and lift gas injection rates.

 Inject gas as deep as possible.
 Determine depth and number of unloading valves. Check if design has sufficient
flexibility to handle current and future well conditions (increased WC, declining
reservoir pressure and PI etc.)
 Produce at stable conditions without "heading" and avoiding pressure surges that could
result in multipoint injection and cause pressure fluctuations large enough to be
disruptive to surface facilities.
 Optimize in terms of revenue by accounting for cost of produced water disposal and lift
gas compression.

The biggest obstacle for performing proper gas lift on Gyda is the capacity of the compressor.
It currently gives 56 600 Sm3/day with an injection pressure of 165 bar. There is much more
gas available, but for a full-scale gas lift campaign a bigger compressor has to be purchased.

For this study, full capacity of 56 600 Sm3/day is assigned to each well. This is assumed to be
a realistic scenario for a full-scale gas lift campaign, with a proper compressor.

Studies done by TENAS has shown that the older Gyda casings, like the ones in A-19 and A-
26, which was drilled in the early nineties, cannot handle the pressure from gas lift injection
lower than 2600 m TVD. This implies that maximum injection depths in these wells will be at
2830 and 2650 m MD.

When designing a gas lift system with more than one well, one should optimize the allocation
of available gas, so that the total production is maximized. The developers of PROSPER has
made a software for this (GAP), but this is not the scope of this study and has not been done.

71 | P a g e
 10.1. Modelling A-19 and A-26 with gas lift
When modelling a gas lift well a number of parameters have to be entered into the system.
Figure 46 shows the gas lift design menu. The gas available for lifting has the following
characteristics:

Table 9.1, Lift gas data

Gas lift gas gravity, sp. gravity 0,784

Mole percent H2S, % 0
Mole percent CO2, % 2,76
Mole percent N2, % 0,98

Figure 46; gas lift design menu.

The operating injection pressure is set to 165 bar. Desired dP across valves, 10 bars, is entered
to ensure well and gas injection system stability. Minimum spacing between valves is set to 60

72 | P a g e
m. Sea water is assumed as the load fluid before gas lift start, which result in a static gradient
of 0,1 bar/m. Also, maximum injection depth for each well is set.

The most used valve type by TENAS is the casing sensitive, which is also chosen here. Valve
settings is chosen to “Pvc = Gas Pressure”. Then PROSPER sets valve dome pressure to
balance casing pressure at depth. Unloading valves will close when the casing pressure drops
below this value.

For this study a “Camco BK Normal” valve is chosen from the PROSPER database. The
software calculates which port sizes that will generate optimal production. A valve from
another manufacturer would maybe require different port sizes, but PROSPER still calculates
the same optimal production. Therefore, the choice of type is not that important as long as it is
casing sensitive.

Now PROSPER is able to calculate a gas lift performance curve. Figure 47 and 48 show the
curves for A-19 and A-26 with the May 2010 conditions (table 7 and 8).

Figure 47; Gas lift performance curve for A-19

73 | P a g e
 Figure 48; Gas lift performance curve for A-26

The performance curves give us a plot of oil produced versus the gas injected. The injection
gas rate that gives the highest production rate can be found, although that might not be the
optimum point of injection in terms of revenue. That point is where the incremental additional
cost of compressing gas equals the incremental revenue of the additional oil produced. The
economic optimum gas injection rate is often found to the left of the maximum production rate
in such a curve. When looking at the curves it is clear that none of the wells will reach
maximum production with an injection rate of 56 600 Sm3/day. A-26 is producing in a steeper
part of its curve, but an increased injection in this well might not increase the total production.
A-19 produces larger oil volumes, so even if we inject more gas into A-26, and get a larger
percentual increase of out of this well, A-19 will give more oil with the same gas injected. This
shows the importance of a full system analysis.

Looking at the curve from A-19 one can see that it starts from approximately 2380 bbl/day.
This coincides with the result from the base model where A-19 produced 2381 bbl/day with
natural flow. The same goes for A-26.

74 | P a g e
 10.2. Positioning of valves
Valve spacing is not affected by the choice of unloading method (casing or tubing sensitive),
but of whether the well IPR is used for calculating the unloading rate or not. When designing
the valve system PROSPER can be set to check whether the solution rate is achievable with
respect to the IPR. IF necessary, the design rate is reduced and the spacing calculation is
repeated. Figures 49 and 50 shows the result of valve spacing design for A-19 and A-26.

Figure 49; valve spacing for A-19

75 | P a g e
 Figure 50; valve spacing for A-26

The valve spacing calculations is done in the following way:

For the tubing, with the designed injected gas rate, a pressure traverse is calculated from the
surface and downwards using the gas lifted flowing gradient (blue line). A similar plot is made
for the casing pressure (right red line).

The injection depth (orifice valve) is the depth at which the flowing tubing pressure equals the
casing pressure gradient less the designed pressure loss across the orifice. However, injection
depth is often limited by well design, for example by a production packer or weak casings like
on Gyda.

The shallowest unloading valve is placed at the depth that balances the tubing load fluid
pressure (left red line) with the casing pressure at that depth. Further unloading valves are
placed by traversing down like this between the casing and gas lifted tubing pressure gradient
lines.

Valves are placed ever deeper until the inter valve spacing equals a pre-set minimum, or the
maximum injection depth has been reached.

76 | P a g e
Once the first design is complete, the software can re-calculate the flowing tubing gradient
using the current operating valve depth. This was not necessary for A-19 and A-26 because
both wells were able to inject at the preset maximum injection depth.

10.3. Results
The result from the gas lift design is given in table 10:

Table 10; Results from gas lift design

Result / Well A-19 A-26

Depth unloading valve, m MD 1828 1714
Unloading valve port size, 64ths inch 12 14
Depth operating valve, m MD 2830 2650
Operating valve port size, 64ths inch 14 14
Oil rate with gas lift, bbl/day 2988 948
Oil rate without gas lift, bbl/day 2381 234

The results show that both wells get a significant increase in production from gas injection.
Both wells only needed one unloading valve. This is a combination of the operating valve
setting depth, gas lift injection pressure and load fluid density. For example if the wells did not
have a depth constrain and was displaced to a 800 kg/m3 fluid instead of seawater when starting
the injection, the operating valve could be set deeper and we would see an increased production.
This will be discussed in the next section.

10.4. Sensitivities of injection depth

To see the effect of injection depth a sensitivity analysis was run on this. PROSPER allows the
injection depth to be a variable, and calculates production for each depth entered. Table 11
shows at which depths the analysis was run and the corresponding results.

77 | P a g e
 Table 11; injection depth analysis on A-19 and A-26

A-19 A-26
100 m MD 2392 bbl/day 100 m MD 309 bbl/day
500 m MD 2435 bbl/day 500 m MD 492 bbl/day
1200 m MD 2597 bbl/day 1200 m MD 721 bbl/day
2000 m MD 2810 bbl/day 2000 m MD 876 bbl/day
2830 m MD 2988 bbl/day 2650 m MD 948 bbl/day
3000 m MD 3021 bbl/day 3000 m MD 979 bbl/day
4000 m MD 3178 bbl/day 3850 m MD 1033 bbl/day

Figure 51; injection depth analysis on A-19. The figure shows how the VLP curve is moved as
a function of injection depth.

78 | P a g e
Figure 52; injection depth analysis on A-26. The figure shows how the VLP curve is moved
as a function of injection depth.

The results show that a deeper setting depth of the operating valve gives an increased
production. However, the results also show that the effect off moving the valve to the bottom
of the 9 5/8” casing, at approximately 3000 m MD in both wells, has a minimal effect (See
completion drawings). There is a bigger effect of moving the valves down to the bottom of the
tubing inside the 7” liner (4000 m and 3850 m MD), which gives 190 and 85 bbl/day more.
This has not been done on Gyda but has been tried with success on another TENAS installation,
Varg. This could be a possibility for newer Gyda wells with more solid casings than A-19 and
A-26.

11. ESP DESIGN

Compared to the gas lift part of this study, an ESP system is not designed in PROSPER. A
study and design done by Baker Hughes Centrilift has provided the system data given in table
12. Upper and lower systems are the same in both wells.

79 | P a g e
 Table 12; ESP system data [19]

Equipment A-19 and A-26

Pump Centrilift 562 P110
Operating range 7000 – 13200 bbl/day (60Hz)
Motor Centrilift 562 450 HP
Cable Grade 1 copper
# Stages 84 for A-19, 60 for A-26

The 562 (Figure 10.1) series is designed for use in 7 inch wells where large production volumes
are required and in 9 5/8 inch casings when space is limited. The outside diameter of the pump
housing is 5,62 inches, but the pump is capable of operating in 7 inch wells when the casing
weight is 26 lbs/ft or lower. All pump models in this series have a 1 3/16 inch shaft, which
allows for a maximum horsepower rating of 1250 at 60 Hz. The 562 series comes standard with
carbon steel metallurgy but certain configurations can be supplied in corrosion resistant alloys
or with optional “Monel coating” when installed in corrosive environments.

As described in section 7.6. it is planned for a dual ESP and sealed shrouds solution. The units
will be put inside the 9 5/8” casing as close to the 7” liner as possible. For this study they are
set at a depth of 3050 m MD. The pump is planned to be operated in the range between 50 and
60 Hz.

80 | P a g e
 Figure 53; Centrilift 562 P110 [19]

11.1. Modelling A-19 and A-26 with ESP

Coefficients (appendix) for head and horsepower received by Centrilift is put into the
PROSPER database. With these coefficients the software calculates pump curves so that one
can simulate for any condition (Figure 54).

81 | P a g e
 Figure 54; pump curves for the Centrilift 562 P110

From the figure we see that the minimum operating range, running at 60 Hz, is 7000 bbl/day
and maximum is 13200 bbl/day. The perfect operating point is at the best efficiency line. This
is where the efficiency curve for the pump peaks (see section 7.1.).

There are two options when modelling an ESP well. One is to enter the design menu and let
PROSPER design the ESP scenario, and choose a pump from the options that the software
suggests. The second option is to enter the pump data directly when type of pump and motor
is already decided, like in this case.

The data is put into the menu in figure 55. Pump, motor and cable are selected from the
database.

82 | P a g e
 Figure 55; ESP input in PROSPER (A-19 as an example)

11.2. Results
Given the May 2010 conditions and ESP data PROSPER calculates production in the ESP
wells:

Table 13; Results from ESP simulation

Result / Well A-19 A-26

Oil rate with ESP, bbl/day 4619 1584
Total liquid rate ESP, bbl/day 8248 8405
Oil rate with gas lift, bbl/day 2988 948
Oil rate without artificial lift, bbl/day 2381 234

The wells show a significant increase in production compared both to the gas lift case and base
case.

83 | P a g e
Both wells have a total liquid rate which lies well within the operating range of the pump,
between the minimum and best efficiency line. This means that there is still a good capacity
for handling more fluid.

Figure 56; A-19 ESP system

For pumps in PROSPER e.g. ESPs, system analysis is conducted at the pump discharge (outlet
of pump). Figure 56 has the following explanation:

The green curve is the IPR pressures referred to the bottom of the well.

The blue curve is the pump discharge pressure. This is the pump intake pressure corrected for
dP added by pump.

The red curve is VLP from top of well to pump discharge (and not bottom of well). The solution
rate is obtained by the intersection of pump discharge pressure and the VLP.

12. PRODUCTION FORECAST

Based on an Eclipse reservoir simulation a sensitivity analysis is made on water cut and
reservoir pressure using PROSPER. PI and GOR are assumed constant. For other assumptions
on water cut and GOR see section 9.2.2.

84 | P a g e
Exact reservoir pressure is difficult to estimate. The reservoir simulation gives flowing and
static BHP, however this is only valid for the nearby wellbore area. The pressure that
PROSPER need is the pressure far out in the reservoir pushing the fluid into the low pressure
area near the wellbore created by the artificial lift. This pressure difference is what makes the
drawdown.

The initial reservoir pressures in 2010 are based on well tests and reservoir simulation of the
“base case” where there is no artificial lift. Both wells are set to 400 bar, which is a conservative
estimate.

Field simulations show that the ESPs will draw down the average pressure of the field with 50
bar, and this is what the reservoir pressure prediction is based on, not the BHP. Table 15 shows
how the reservoir pressure decrease. The reason why the reservoir pressure will not decrease
to BHP is that there is pressure support from two injectors A-9 and A-28. This can also be seen
from the constant reservoir pressure in the A-19 “base case”.

The gas lift is assumed to draw down the reservoir pressure 10 percent of what the ESPs does
(see table 16).

The “secondary effect” of ESPs described in section 7.7 is not accounted for in the reservoir
simulation this study is based on. Therefore the oil production forecast of the ESP wells will
be conservative if this effect is valid.

12.1. Well A-19

Table 14, 15 and 16 show the result from the sensitivity analysis done from 2010, which is
installation date, to 2019. Figure 57 shows the production forecast comparing the base case,
gas lift and ESP

Table 14; A-19 “base case”

A-19 “base case”

Parameter/year 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
GOR 140 140 140 140 140 140 140 140 140 140
WC 53 55 60 65 67 70 71 72 74 75
Pres 400 400 400 400 400 400 400 400 400 400

85 | P a g e
 Qo, bbl/d 2381 2200 1764 1342 1176 939 865 789 636 562

Table 15; A-19 “ESP case”

A-19 “ESP case”

Parameter/year 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
GOR 140 140 140 140 140 140 140 140 140 140
WC 44 48 55 60 63 67 71 73 76 78
Pres 400 375 360 355 355 355 355 355 355 355
Qo, bbl/d 4619 3929 3179 2717 2469 2140 1894 1732 1494 1260
Tot. liquid, bbl/d 8247 7555 7065 6791 6673 6483 6315 6187 5973 5729

Table 16; A-19 “gas lift case”

A-19 “gas lift case”

Parameter/year 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
GOR 140 140 140 140 140 140 140 140 140 140
WC 53 55 60 65 67 70 71 72 74 75
Pres 400 398 395 395 395 395 395 395 395 395
Qo, bbl/d 2988 2807 2389 2022 1880 1672 1605 1537 1406 1341

Figure 57; Production forecast for A-19. Production for the base case and the two artificial lift
methods are plotted from start date to 2019.

86 | P a g e
The result of the production forecast show that the ESP solution gives a superior production
rate compared to gas lift and the “base case”. From figure 57 it is observed that the gas lift
production rate crosses the ESP production rate in 2019. This is a result of the initially higher
water cut now being lower than in the ESP case. The gas lift scenario also does not see the
same reservoir pressure loss as the pump.

The total liquid rate falls beneath 7000 bbl/day after 2012 in the “ESP case”. This is below the
minimum operating range of the pumps running at 60 Hz. When this happens, the pumps have
to be run at a lower frequency (Fig.54).

When both the pumps fail, there should be a new analysis with the current conditions to see if
another pump design would fit better. After running some years, one would also learn more
about the rates and how the reservoir responds to the pumps.

12.2. Well A-26

Table 17 and 18 show the result from the sensitivity analysis done from May 2010, which is
installation date, to 2019. Figure 58 shows the production forecast. There is no “base case”
scenario because A-26 is considered as dead.

Table 17; A-26 “ESP case”

A-26 “ESP case”

Parameter/year 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
GOR 160 160 160 160 160 160 160 160 160 160
WC 81 84 85 87 88 90 91 92 93 93
Pres 400 375 360 355 355 355 355 355 355 355
Qo, bbl/d 1598 1130 939 741 663 516 448 382 320 320
Tot. liquid, bbl/d 8410 7063 6260 5700 5524 5162 4976 4773 4571 4571

Table 18; A-26 “gas lift case”

A-26 “gas lift case”

Parameter/year 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
GOR 160 160 160 160 160 160 160 160 160 160

87 | P a g e
 WC 86 88 89 90 91 91 92 92 93 93
Pres 400 398 395 395 395 395 395 395 395 395
Qo, bbl/d 948 787 703 631 561 561 493 493 426 426

Figure 58; production forecast for A-26. Production for the base case and the two artificial lift
methods are plotted from start date to 2019.

The results show that also in A-26 ESP give the largest production. However, gas lift shows a
higher production rate already in 2015. Although the difference in production is not as
impressive in A-19, one can still see from figure 58 that the area under the ESP curve is much
larger than for the gas lift curve.

There is a small “base case” production included in 2010. The well dies with a small change in
water cut and reservoir pressure.

The total liquid rate falls beneath 7 000 bbl/d after only two years. If the rate follows this
prediction the pump has to be run at lower frequency or changed.

The reservoir pressure used is conservative, and a pressure of 450 bar instead of 400 bar is
realistic. This would give a liquid rate of over 10 000 bbl/day. This shows that pumps are not
necessarily oversized, but has to handle a wide range of scenarios and production rates.

The reliability of the production forecast is dependent of the reservoir pressure, and mostly the
water cut. A small change in water cut gives a large change in oil production. The reservoir

88 | P a g e
pressure is conservatively estimated, and the water cut trend is based on reliable reservoir
simulations. In A-19 there has not been a major water break-through yet, and the timing of this
breakthrough involves some uncertainty. However, the same reservoir data source is used for
the different scenarios, so the comparison of the two artificial lift methods should be valid.

13. ECONOMICAL EVALUATION

Before making a final decision a thorough economic analysis has to be done. As described in
section 6.2.3, it is the profitability of a project that has to be the final decision criteria. TENAS
is still in the evaluation phase, and a full economic analysis giving the NPV of the projects is
not available yet. The NPV will give the value of a project through its entire lifetime taking
capital costs, operating costs, depreciation and revenues into account.

However, the initial costs of the scenarios are analysed and can give a good indication of the
project magnitude. Table 19 shows the capital cost e.g. the cost until end of installation of each
project. This involves cost of procurement, construction, engineering, administration and
operational cost during installation (rig rate etc.). The numbers do not include company costs
such as company personnel, helicopter, catering etc.

Table 19; capital cost of projects.

Project Total cost (1000 NOK)

ESP pilot, A-19 and A26 284 000
ESP full field project, 10 wells 1 420 000
Full field gas lift project 709 000

There has not been made any economical estimate on gas lift just for A-19 and A-26. The plan
is either full field or nothing. The ESP full field project estimate is the pilot project times five.
This is done so simple because it is too early to say what a full field project would imply. Ten
wells producing with ESPs would need a new water handling facility etc. Thus, a full field ESP
project can be more expensive.

The relatively high expenses of full field gas lift project are caused by a new compressor having
to be purchased, and a lot of the wells would need a full workover due to missing side pockets.

89 | P a g e
The biggest cost for gas lift is at installation, while ESPs carries great operational expenses due
to their limited lifetime. A workover for two wells with ESPs is estimated to 60 days. This
means that an ESP solution has a great cost attached every two years. The operational costs are
estimated to about 75% of the total costs of a pump change out. The loss of production during
this time also needs to be accounted for.

The cumulative oil production can be multiplied with a factor of (1-2/24), where 2 is the
downtime and 24 is the expected lifetime in months. Compared, the cost of a gas lift valve
change out by wireline is minor. This operation will only take a day or two.

14. CONCLUSIONS
Through consideration of the production profile, desired rate and advantages / disadvantages
TENAS has decided for gas lift and ESP to be the most suitable artificial lift methods on Gyda.

Gas lifting is a simple, well tried method and has been proved efficient on Gyda before.
Installation and change of gas lift valves are done by wireline, but since the A- 19 and A-26 do
not contain side pockets, a full workover is required at first time installation.

There is a positive effect of setting the valves deeper. When the compressor outlet pressure is
limited, the fluid density in the well is important. The valves can be set deeper with a less dense
fluid, and this can also make the difference in number of unloading valves needed.

A new compressor has to be purchased to reach the injection rates used in this study. This will
be one of the biggest costs of an artificial lift campaign based on gas lift.

Implementation of ESPs carries greater risk because of the complexity of the equipment and
limited lifetime. When ESPs fail this require a full workover, which is costly mainly because
of the required rig operation compared to a wireline operation. However, there are design
choices and running procedures that will extend the lifetime. SAGD technology is used to cope
with the high reservoir temperatures on Gyda. Monitoring production and the pump during
operation is crucial to achieve extended lifetime. Sand production and scale is two of the
biggest risks.

Expected lifetime of the dual ESP design on Gyda is 2 years.

Both gas lift and ESP give a large increase in production compared to the base case, but ESP
is superior to gas lift in both A-19 and A-26. It is reason to believe that the same difference

90 | P a g e
would be seen in a full field artificial lift campaign. In this study the so called “ESP secondary
effect” is not accounted for, this can increase production as well as the recovery factor for the
field. So, from a production point of view the ESPs is by far the best choice.

The difference in oil production between ESP and gas lift is largest in A-19. Both wells show
basically the same total liquid production initially. However, as water cut increases and
reservoir pressure decreases in both wells, A-26 cannot follow A-19’s production trend with
ESPs. The reason seems to be the lower number of stages in A-26, this shows that a re-
evaluation of the design is needed when it is time for changing the pumps.

The water cut behavior is the biggest factor for change in the oil production. To date A-19 has
not seen any significant water breakthrough, but this is expected to happen soon. A water
breakthrough is accounted for in the simulations done in this study.

With lower water cut, A-19 will produce at a larger rate than predicted.

Since ESPs have never been run on Gyda before it is difficult to tell how the reservoir will
react. Reservoir pressure and water cut is difficult to predict and the pumps must be able to
handle a wide range of fluid ranges and properties. Baker Hughes Centrilift’s design seems to
be valid for the conditions predicted for May 2010, where production lies to the left of the best
efficiency point (Figure 10.2). At this point they have extra capacity, and can also run at a
lower rate.

According to the production forecast done in this study, there is a possibility for the pumps to
fall beneath their production design limit in both A-26 and A-19 after some years. This again
shows the importance of a re-evaluation of the design when a pump fails.

Comparing cost and production potential of the artificial lift methods, ESPs are the best choice.
The pilot project with A-19 and A-26 would return invested capital in less than a year due to
its high production potential compared to the base case.

But before a final decision is made an economic analysis of each project’s lifetime should be
carried out. The ESP projects will generate higher costs later in life than the gas lift project. An
NPV evaluation will account for all costs and depreciation of each project.

91 | P a g e
REFERENCES
1. Michael Golan and Cutis H. Whitson, Well Performance, Prentice-Hall Inc., 1985.
2. Gabor Takacs, Gas Lift Manual, PennWell Corporation, 2005.
3. Dake, L., Fundamentals of Reservoir Engineering, Elsvevier Science B.V, 2002.
4. Clegg, J.D., editor. Production Operations Handbook. SPE, USA. 2007.
5. Rune W. Time, Two phase Flow in Pipelines, Course compendium, University of
Stavanger, 2007.
6. PROSPER v.10 User Manual, Petroleum Experts. 2008.
7. Clegg, J.D., Bucaram, S.M., Heln Jr., N.W., Recommendations and Comparisons for
selecting Artificial-Lift Methods, SPE 00024834, 1993.
8. Lloyd R. Heinze, Herald W. Winkler, James F. Lea, Decision Tree for Selection of
Artificial Lift Method, SPE 29510, 1996.
9. Gaviara, F., Santos, R., Rivas, O., Luy, Y., Pushing the Boundaries of Artificial Lift
Applications: SAGD ESP installations in Canada, SPE 110103, 2007.
10. Vogel, J.V., Inflow Performance Relationships for Solution-Gas Drive Wells, SPE
1476-PA, 1968.
11. Kilvington, L.J., Thomson, J.Y., Brown, J.K., Electric-Submersible-Pumping Success
in the Beatrice Field, North Sea, SPE 16638-PA, 1989.
12. Meihack, S., Performance of Jet Pump in the Gyda field, Internal document of
Talisman Energy Norge AS, Stavanger, 1997.
13. Zerrouki, T., Paul, H., Monkman, J., Yme: Expected ESP Run Life, Internal
document of Talisman Energy Norge AS, Stavanger, 2006.
14. Saepudin, D., Soewono, E., Sidarto, K.A., Gunawan, A.Y., Siregar, S., Pudjo, S., An
Investigation on Gas Lift Performance Curve in an Oil- Producing Well, Hindawi
Publishing Corporation, International Journal of Mathematics and Mathematical
Sciences, Volume 2007, Article ID 81519.
15. Fleshman, R., Harryson, O.L., Artificial Lift for High-Volume production, Oilfield
Review, spring 1999.
16. Jadid, M.B., Lyngholm, A., Opsal, M., Vasper, A., The Pressure’s On: Innovations in
Gas Lift, 2006, Volume 18.
17. Basic artificial lift, Canadian Oilwell Systems Company Ltd, available at:
http://www.coscoesp.com
18. Baker Hughes Oil Tools, Gas Lift catalog, available at:
http://www.bakerhughesdirect.com

92 | P a g e
 19. http://www.bakerhughes.com
20. http://www.ptc-norway.com
21. http://www.petroleumexperts.com
22. http://www.pumptools.co.uk/
23. http://www.npd.no
24. http://www.slb.no

ABBREVIATIONS
ADV Automatic Diverter Valve

BEP Best Efficiency Point

BHP Break Horsepower

EOR Enhanced Oil Recovery

EOS Equation of State

ESP Electrical Submersible Pump

FBHP Flowing Bottom Hole Pressure

GLR Gas-to-Liquid Ratio

GLV Gas Lift Valve

GOR Gas oil Ratio

HP Horse Power

IPO Injection Pressure Operated

IPR Inflow Performance Relationship

KOT Kickover-tool

MD Measured Depth

MTTF Mean Time To failure

NPV Net Present Value

OWC Oil Water Contact

93 | P a g e
 PI Productivity Index

PPO Pressure Operated Valves

RIH Running in Hole

SAGD Steam Assisted Gravity Drainage

SG Specific Gravity

TENAS Talisman Energy Norge AS

TVD True Vertical Depth

VLP Vertical Lift Performance

WC Water Cut

94 | P a g e

View publication stats