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

net/publication/340412472

Petroleum Production Systems - Compendium

Book · June 2021

CITATIONS READS
2 9,013

1 author:

Milan Stanko
Norwegian University of Science and Technology
72 PUBLICATIONS 258 CITATIONS

SEE PROFILE

All content following this page was uploaded by Milan Stanko on 26 October 2022.

The user has requested enhancement of the downloaded file.

Petroleum Production Systems

Compendium
Prof. Milan Stanko
Trondheim, Norway
© 2022, Stanko.
Version 1.7.3 (October‐2022)
PREFACE
These notes address, hopefully in a simple manner, a variety of topics on production performance of oil and
gas fields.
The notes are given as compendium for the master‐level courses Field Development and Operations
(TPG4230) and Production Wells (TPG4245) taught at the Department of Geoscience and Petroleum of the
Norwegian University of Science and Technology (NTNU) in Trondheim, Norway. The course TPG4230 was
designed in 2007 by Prof. Michael Golan and teaches and integrates a variety of multi‐disciplinary petroleum
engineering topics used in the development and management of hydrocarbon reservoirs and fields.
The lectures of the courses are video‐recorded and are available on my YouTube channel, under the following
link1. Each lecture has in the description links to my handwritten notes and exercise class files that were
discussed. The videos of the production wells course are available in this link2.
These notes will be updated often, and more material will be added with time. Be aware that references might
be incomplete. If you have any comments or find errors, I appreciate you sending me an email at
milan.stanko(at)ntnu.no. Equation usage is intentionally reduced to a minimum as expressions are usually
provided in class or are available in other sources.
I appreciate and acknowledge the contribution, corrections, time and support of Prof. Michael Golan and Prof.
Curtis Whitson. Many of the ideas presented in the document are based on their work and their way of
thinking.
I appreciate the help and contributions of Ruben Ensalzado regarding document formatting, editing, re‐writing
numerous equations and general quality control.

Prof. Milan Stanko

1
https://www.youtube.com/channel/UCWMfsCe1NQMgx4UZWrVvFgA
2
https://www.youtube.com/playlist?list=PLXfmJjG2tXbpjx6bezD4famP9YtVFEXqw

3
CONTENTS
Preface 3
Contents 4
List of tables 8
List of figures 9
1. Field Performance 17
1.1. Reservoir 17
1.2. Production system (surface network) 19
1.3. Coupling reservoir models and models of the production system 20
1.4. Production potential 22
1.5. Production scheduling 23
1.6. Relationship between production potential and cumulative production 25
1.7. Production scheduling and planning using production potential curve 31
1.8. Applicability of the production potential concept in real fields and multi‐well production systems 36
References 37
2. Flow Performance in Production Systems 38
2.1. Inflow performance relationship 40
2.1.1. Undersaturated, vertical oil well 43
2.1.2. Vertical gas well 44
2.1.3. Saturated, vertical oil well 45
2.1.4. Composite IPR: Both undersaturated and saturated oil 48
2.1.5. Flow of associated products in an oil well: gas and water 48
2.1.6. IPR and water or gas coning 49
2.1.7. IPRs generated with reservoir simulator 50
2.2. Available and required pressure function 50
COMPLETION BITE: Tubulars 54
2.3. Flow equilibrium in production systems 56
2.3.1. Single well production system 56
2.3.2. Operational envelope: choke 59
2.3.3. Booster 63
2.3.4. Operational envelope: electric submersible pump 70
2.3.5. Operational envelope: dynamic gas compressor 72
2.3.6. Operational envelope: jet pump 75
2.4. Flow equilibrium in production networks 77
2.4.1. Solving network hydraulic equilibrium fixing well rates 80
2.4.2. Downhole networks 81
COMPLETION BITE: Sliding sleeve 82
Shifting procedure 84
References 85
3. Production Optimization 86
3.1. Optimizing a production system 86
3.1.1. Case 1: Gas‐lifted wells 88

4
 COMPLETION BITE: gas‐lift valve 90
3.1.2. Case 2: Two gas wells equipped with wellhead chokes 94
3.1.3. Case 3: Two ESP‐lifted wells 97
3.2. Issues hindering the industrial scale adoption of model‐based production optimization 103
3.2.1. Foreign from the field’s reality 103
3.2.2. Models uncertainty 103
3.2.3. Non‐sustainability of the proposed solutions 103
References 104
4. Fluid Behavior Treatment in Oil and Gas Production Systems 105
4.1. The Black Oil Model 105
4.2. Variation of BO properties with temperature 110
4.3. Variation of BO properties with composition 112
4.4. BO correlations 117
4.5. BO properties in production calculations 118
4.6. Estimation of a new composition when the well GOR changes 119
4.7. Water‐related properties 119
References 122
5. The Field Development Process 123
5.1. Business case identification 126
5.1.1. Reserve estimation using probabilistic analysis 126
5.2. Project Planning 130
5.2.1. Feasibility studies 130
5.2.2. Concept planning (leading to dg2) 130
5.2.3. Field production profile and economic value 131
5.2.4. Pre‐Engineering (leading to DG3) 140
5.3. Project Execution 140
5.3.1. Detailed engineering, construction, testing and startup 140
5.4. Operations 140
5.5. Decommissioning and abandonment 140
References 142
6. Offshore Structures for Oil and Gas Production 143
6.1. Selection of proper marine structure 145
6.1.1. Water depth 145
6.1.2. Location of the Christmas tree 145
COMPLETION BITE: Wellhead architecture 145
Safety strategy for wells 148
6.1.3. Oil Storage 150
6.1.4. Marine loads on the offshore structure 150
6.1. Treatment of wind, waves and currents 152
References 157
7. Flow Assurance Management in Production Systems 158
7.1. Hydrates 159
7.1.1. Consequences 160

5
 7.1.2. Management 160
7.2. Slugging 163
7.2.1. Consequences 164
7.2.2. Management 164
7.3. Scaling 165
7.3.1. Consequences 165
7.3.2. Management 165
7.4. Erosion 166
7.4.1. Consequences 166
7.4.2. Management 166
7.5. Corrosion 167
7.5.1. Consequences 167
7.5.2. Management 167
7.6. Wax Deposition 168
7.6.1. Consequences 169
7.6.2. Management 170
7.7. Oil‐Water Emulsions 171
7.7.1. Consequences 171
7.7.2. Management 173
7.1. Summary table 173
7.1. About chemical injection 174
References 176
8. Heat Transfer for flow in conduits 177
8.1. Order of magnitude analysis on the specific energy terms of a stream 177
8.1.1. Comparison between the specific kinetic and potential energy terms 178
8.1.2. Comparison between the specific enthalpy and potential energy terms 181
8.2. Heat Transfer with the environment 182
8.2.1. Case 1. Subsea pipeline 183
8.2.1. Case 2. Heat transfer in wellbore 187
8.3. Behavior of specific enthalpy of oil and gas versus pressure and temperature for multi‐component
hydrocarbon mixtures 192
8.4. Procedure to estimate temperature drop in a conduit 194
References 195
9. Gas well liquid loading 196
9.1. Liquid loading maps 199
9.2. “Transientness” of liquid loading 201
References 203
Appendices 204
A. The Tubing Rate Equation in Vertical and Deviated Gas‐Wells 205
Derivation from first principles (pure SI system) 205
Pressure equation in practical field units (Metric) 207
Fetkovich Rate Equation 209
References 212

6
B. Choke Equations 213
Undersaturated oil flow 213
Dry gas flow 214
Oil‐Gas‐Water mixture 216
C. Temperature Drop in Conduit for Liquid Flow 219
General expression 219
Derivation for liquids 219
With variable ambient temperature 221
D. Derivation of multiphase flow expressions 222
Relationship between liquid holdup (HL), slip ratio (S) and non‐slip liquid volume fraction (L) 222
Relationship between holdup (HL), slip ratio (S) and quality (x) 222
Holdup average mixture density (ρm) 223
Effective momentum density 223
Kinetic energy‐average mixture density 224
E. Oil & Gas Processing Diagrams 226
F. Derivation of local mass and volume fractions of oil gas and water as a function of black oil properties
228
Gas, oil and water mass fractions 228
Gas, oil and water volume fractions 231
G. Derivation of the expression of field producing gas‐oil ratio 234
H. Gas lift optimization 235
I. Analytical expression of revenue NPV considering a linear dimensionless production potential, continuous
discounting and oil price varying linearly in time. Case study: oil offshore field. 238
J. Some style comments for technical communication (paraphrasing the notes of M. Standing and M. Golan)
240

7
LIST OF TABLES
Table 2‐1. Time required to pss for a gas reservoir with the characteristics and using Eq. 2‐1. 42
Table 2‐2. Required boosting power and outlet temperatures when boosting 7 kg/s of several fluids from 50
°C, 50 bara to 100 bara, assuming adiabatic efficiency of 0.6 68
Table 3‐1. Polynomial coefficients 92
Table 4‐1. BO parameters 106
Table 4‐2. Selected correlations for BO parameters 117
Table 6‐1. Qualitative storage capacity of common offshore structures 150
Table 7‐1. summary table of flow assurance issues: causes, potential consequences, prevention and solution
measures and tools available for analysis 174
Table 8‐1. Upper bounds on temperature and pressure spatial gradients. 178

8
LIST OF FIGURES
Figure 1‐1. (a) Tank analogy of a (b) reservoir system 18
Figure 1‐2. Graphical depiction of the material balance approach 18
Figure 1‐3. IPR curve 20
Figure 1‐4. Time step of an explicit coupling scheme between a material balance model and a model of the
production system to predict production profile 21
Figure 1‐5. Explicit coupling between a reservoir simulation and a model of the production system 21
Figure 1‐6. (b) Well potential calculation vs. (a) Production potential calculation 22
Figure 1‐7. Production potential behavior vs. time when a production enhancement modification is performed
in the system 23
Figure 1‐8. Plateau production mode 24
Figure 1‐9. Production profile obtained when operating in decline mode 24
Figure 1‐10. Production rate behavior vs cumulative production for open choke and constant rate 25
Figure 1‐11. Changes of IPR with cumulative production 25
Figure 1‐12. Production rate behavior vs cumulative production for open choke showing the region of feasible
rates 26
Figure 1‐13. Field production potential vs cumulative production for dry gas reservoir with standalone wells
28
Figure 1‐14. Dimensionless field production potential vs recovery factor for dry gas reservoir with standalone
wells 28
Figure 1‐15. Dimensionless field production potential vs recovery factor for dry gas reservoir with standalone
wells, sensitivity study on system properties 29
Figure 1‐16. Dimensionless field production potential vs recovery factor for dry gas reservoir with standalone
wells, network wells and considering IPR only. 30
Figure 1‐17. Dimensionless field production potential vs recovery factor for several production systems. 31
Figure 1‐18. Plateau mode production 33
Figure 1‐19. Example case: 2 standalone wells 34
Figure 1‐20. 4 different alternatives to produce the two wells system in plateau mode 36
Figure 2‐1. Simplified level and pressure control system in a separator 38
Figure 2‐2. Layout of two production systems 38
Figure 2‐3. IPR curve 39
Figure 2‐4. Production network with two wells 40
Figure 2‐5. Cross section of a vertical well depicting the coordinate system to plot pressure versus radius 41
Figure 2‐6. Evolution of pressure across the reservoir with time when put on production 41
Figure 2‐7. IPR predicted by Eq. 2. for undersaturated oil well and different reservoir pressures 44

9
Figure 2‐8. Graphic illustration of the process to estimate IPR with a reservoir simulator according to Astutik
(2012) 50
Figure 2‐9. Pipe segment 51
Figure 2‐10. Available pressure at pipe outlet for different flow rates and fixed inlet pressure 51
Figure 2‐11. Required pressure at pipe inlet for different flow rates and fixed outlet pressure 51
Figure 2‐12. Available wellhead pressure vs produced rate 52
Figure 2‐13. Available wellhead pressure with choke included vs produced rate 52
Figure 2‐14. Required flowing bottom‐hole pressure curve vs. produced rate 53
Figure 2‐15. Schematic representation of the mixture density variation with slip between gas and liquid
velocities 53
Figure 2‐16. Typical flow patterns along a wellbore as pressure and temperature decrease 54
Figure 2‐17. Two joints of tubing joined by a coupling, or an integrated joint 55
Figure 2‐18. Equilibrium flow rate of the system calculated by intersecting the available pressure curve
calculated from reservoir and the required pressure curve from separator 57
Figure 2‐19. Equilibrium flow rate of the system for: fully open choke and 75% open choke 57
Figure 2‐20. Equilibrium analysis excluding the wellhead choke to estimate choke pressure drop to achieve a
specific flow rate 58
Figure 2‐21. Equilibrium analysis excluding the ESP to estimate ESP pressure boost to achieve a specific flow
rate 58
Figure 2‐22. Equilibrium analysis excluding the choke to estimate choke pressure drop to achieve a specific
flow rate for different times 59
Figure 2‐23.Performance curve of a choke with fixed opening 59
Figure 2‐24. Positive (fixed) choke in critical regime (sonic velocity reached at the throat) 60
Figure 2‐25. Pressure along the axis of a bean choke 60
Figure 2‐26. Performance curve of a choke with fixed opening 61
Figure 2‐27. Performance curve of an adjustable choke for several choke openings 61
Figure 2‐28. Flow rate across different types of adjustable chokes with a fixed pressure drop and inlet pressure
62
Figure 2‐29. Wellhead equilibrium analysis for choke design for two depletion states 62
Figure 2‐30. Adjustable choke performance curve for different choke openings and two inlet pressures 63
Figure 2‐31. Pressure‐enthalpy diagram for methane depicting an isentropic compression process and a real
compression process 65
Figure 2‐32. Pressure‐enthalpy diagram for a dead oil depicting lines of iso‐temperature and iso‐entropy 65
Figure 2‐33. Pressure‐enthalpy diagram for an oil with GOR = 150 depicting lines of iso‐temperature and iso‐
entropy 66

10
Figure 2‐34. Pressure‐enthalpy diagram for an oil with GOR=500 depicting lines of iso‐temperature and iso‐
entropy 67
Figure 2‐35. Pressure‐enthalpy diagram for an oil with GOR=10 000 depicting lines of iso‐temperature and iso‐
entropy 67
Figure 2‐36. Pressure‐enthalpy diagram for methane depicting lines of iso‐temperature and iso‐entropy 68
Figure 2‐37. Diagram of hs versus mass flow for a booster operating at constant power of 1.4 MW and
assuming a constant adiabatic efficiency of 1 69
Figure 2‐38. Pump performance curve, delta pressure vs local flow rate 70
Figure 2‐39. Pump performance curve operating with water or with an oil of 200 cp. Predicted with the method
described in the standard ANSI/HI 9.6.7‐2010 71
Figure 2‐40. ESP equilibrium analysis for ESP design for two depletion states 72
Figure 2‐41. ESP performance curve with operating points overimposed 72
Figure 2‐42. Pressure‐enthalpy diagram for methane depicting an isentropic compression process and the
output of a real compression process 73
Figure 2‐43. Performance map of a gas compressor 74
Figure 2‐44. Simplified schematic of a jet pump 76
Figure 2‐45. Performance plot of a jet pump (taken from Beg and Sarshar[2‐7]) 76
Figure 2‐46. Production network with 3 wells. Available junction pressure curve for three wells and required
junction pressure curve for the pipeline 77
Figure 2‐47. Depiction of the production network model as a mathematical function 78
Figure 2‐48. Production network with 2 wells 78
Figure 2‐49. Well flow rate solutions for no choke, well 1 closed, and well 2 closed 79
Figure 2‐50. Well flow rate domain solution for the production system with 2 wells 79
Figure 2‐51. Well flow rate domain solution for the production system with 3 wells 80
Figure 2‐52. Well flow rate domain solution for the production system with 2 wells 81
Figure 2‐53. Horizontal wellbore with several sections delimited by packers 82
Figure 2‐54. Equivalent line diagram representing a sectioned horizontal wellbore 82
Figure 2‐55. Generic sliding sleeve configuration 83
Figure 2‐56. Details of the locking mechanism of the sleeve 83
Figure 2‐57. Details of the locking fingers on the sleeve that retract and expand when reciprocated axially
inside the sleeve 83
Figure 2‐58. Shifting sequence of a sliding sleeve using slickline 84
Figure 3‐1. Data assimilation process for the network model (adapted from Barros et al, 2015[3‐3]) 87
Figure 3‐2. Equilibrium flow rate of the system for: fully open choke, 75%, 50% and 25% open choke 88
Figure 3‐3. Natural equilibrium point calculated for well with no gas lift injection and with gas injection 89

11
Figure 3‐4. Natural equilibrium points calculated for different amounts of gas lift injected 89
Figure 3‐5. Gas‐lift performance relationship 89
Figure 3‐6. Mandrel types used to deploy gas‐lift valves 90
Figure 3‐7. Locking process of the gas lift valve in the mandrel pocket 91
Figure 3‐8. Sequence to retrieve a gas‐lift valve from the mandrel pocket 91
Figure 3‐9. Colormap and contour lines of total oil production as a function of lift‐gas injected in wells 1 and 2
92
Figure 3‐10. Color map and contour lines of oil production as a function of lift‐gas injected in wells 1 and 2.
Contour lines of total available gas‐lift rate. 93
Figure 3‐11. Production system with two dry gas wells 94
Figure 3‐12. Well flow rate domain solution for the production system with 2 wells 95
Figure 3‐13. Total gas production as a function of well 1 and well 2 rates 95
Figure 3‐14. Total gas production plotted on the feasibility region 96
Figure 3‐15. Well flow rate domain solution for the production system with 2 wells. Well 2 has a higher
deliverability than well 1 96
Figure 3‐16. Total gas production plotted on the feasibility region 97
Figure 3‐17. Two ESP‐lifted wells with common wellhead manifold discharging to a pipeline 98
Figure 3‐18. Total oil production color map for the complete ESP frequency range of wells 1 and 2 98
Figure 3‐19. Two ESP‐lifted wells with common wellhead manifold discharging to a pipeline. 100
Figure 3‐20. Two ESP‐lifted wells with common wellhead manifold discharging to a pipeline. 101
Figure 3‐21. Feasible operating region of a system with two ESP‐lifted wells with common wellhead manifold
discharging to a pipeline. 102
Figure 4‐1. Schematic representation of the flashing of oil and gas at local conditions to standard conditions
106
Figure 4‐2. Schematic of the process to generate BO properties 107
Figure 4‐3. Behavior of BO parameters vs. pressure for a fixed temperature 108
Figure 4‐4. Phase diagram of the hydrocarbon mixture used in Figure 4‐3 108
Figure 4‐5. Behavior of BO parameters vs. pressure for a fixed temperature 109
Figure 4‐6. Phase diagram of the hydrocarbon mixture used in Figure 4‐5 109
Figure 4‐7. Solution gas oil behavior with pressure for three temperatures 110
Figure 4‐8. Phase diagram of the hydrocarbon mixture used in Figure 4‐7 110
Figure 4‐9. Oil volume factor behavior with pressure for three temperatures 111
Figure 4‐10. Gas volume factor behavior with pressure for three temperatures 111
Figure 4‐11. Solution Oil‐gas ratio with pressure for three temperatures 112
Figure 4‐12. Phase diagram of the hydrocarbon mixture used in Figure 4‐12 112

12
Figure 4‐13. Black oil properties estimated for different compositions (GOR). Note that for GOR=1070, the fluid
is not anymore an oil but a gas in undersaturated conditions. 113
Figure 4‐14. Rs and 1/rs vs p computed for several GORs at constant temperature 114
Figure 4‐15. Oil well 115
Figure 4‐16. Variation of the phase envelope with changes in composition (GOR) 115
Figure 4‐17. Rs variation with composition when the new GOR is lower than the original GOR 115
Figure 4‐18. Rs variation with composition when the new GOR is higher than the original GOR 116
Figure 4‐19. Variation of main BO parameters with composition when more gas flows into the wellbore 117
Figure 4‐20. Transformation matrixes to take standard conditions rates to local conditions and vice versa 118
Figure 4‐21. Transformation matrixes to take standard conditions densities to local conditions and vice versa
118
Figure 4‐22. Recombination of source gas and oil to yield stream composition 119
Figure 4‐23. Color map of as a function pf pressure and temperature 120
Figure 4‐24. Color map of as a function of pressure and temperature depicting three possible trajectories
of pressure and temperature in a conduit (departing from an inlet pressure and temperature of 200 bara and
70 °C) 121
Figure 5‐1. Field development timeline and the evolution of the value chain model after decision are made
124
Figure 5‐2. Detailed value chain components 124
Figure 5‐3. Field development process 125
Figure 5‐4. model or simulation with uncertainty in its input parameters 127
Figure 5‐5. probability distribution of initial oil in place calculated with monte carlo simulation and different
number of samples 128
Figure 5‐6. probability distribution of initial oil in place sampled from a normal distribution for different
number of samples 129
Figure 5‐7. Behavior of the net present value of the revenue versus oil plateau rate for two numbers of wells
135
Figure 5‐8. Project NPV, NPVREV , NPVOPEX, and CAPEX+DRILLEX for 12 producer wells versus oil plateau rate.
137
Figure 5‐9. Color contour of NPV versus number of producing wells and field plateau rate 138
Figure 5‐10. Color map of NPV versus number of producing wells and field plateau rate, for three values of
initial oil in place. The blue dot indicates the optimal combination of number of producers and oil plateau rate
139
Figure 5‐11. Color map of NPV versus number of producing wells and field plateau rate, for oil price 50 usd/stb
and 70 usd/stb. 139
Figure 6‐1. Some common marine structures for oil and gas exploitation 144
Figure 6‐2. a) Catenary mooring, b) taut mooring. (Adapted from Chakrabarti[6‐5]) 144

13
Figure 6‐3. Water depth range of the most common offshore structures for hydrocarbon production 145
Figure 6‐4. Deployment of the conductor 146
Figure 6‐5. Run of the surface casing and casing head 146
Figure 6‐6. Details of the pressure port on the casing head to make the pressure test 147
Figure 6‐7. Casing head with the intermediate casing hanged 147
Figure 6‐8. Details of the casing hanger (slips and seals) 147
Figure 6‐9. Installation of the casing spool to the casing head 148
Figure 6‐10. Final configuration of the wellhead 148
Figure 6‐11. Top tension systems for production risers in floating structures (Adapted from Chakrabarti[6‐6])
149
Figure 6‐12. Wind and current loads on an offshore structure 150
Figure 6‐13. Examples of typical movements exhibited by offshore structures 151
Figure 6‐14. Heave RAO of a Sevan FPSO (taken from Saad et al. [6‐7]) 152
Figure 6‐15. Illustrative figure indicating natural periods of some offshore structures and excitation periods of
some environmental loads 152
Figure 6‐16. Wind rose, (Adapted from https://sustainabilityworkshop.autodesk.com/buildings/wind‐rose‐
diagrams) 153
Figure 6‐17. Two‐dimensional random wave time profile 153
Figure 6‐18. Contribution of individual regular waves 154
Figure 6‐19. Wave energy spectrum a) continuous and b) discretized 154
Figure 6‐20. Short term probability density function of wave elevation (a) and height (b) 155
Figure 6‐21. Scatter Diagram of long term wave statistics 155
Figure 6‐22. Pdf and cd of significant wave height for spectral period range 18‐19 s 156
Figure 7‐1. Flow assurance problems and their typical location in the production system 158
Figure 7‐2. A) appearance of a hydrate plug (photo taken from schroeder et al[7‐1] ), b) molecular structure of
a methane hydrate 159
Figure 7‐3. Hydrate formation region 160
Figure 7‐4. Evolution of p and T of the fluid when flowing along the production system 160
Figure 7‐5. Effect of inhibitor injection on the hydrate line 161
Figure 7‐6. Flow schematic of a subsea production system with hydrate inhibitor injection system 162
Figure 7‐7. Details of a subsea distribution unit. 162
Figure 7‐8. Hydrate and scale inhibitor injection system in the X‐mas tree 163
Figure 7‐9. Slug in a pipe section 163
Figure 7‐10. Flow pattern map for a horizontal pipe 164

14
Figure 7‐11. Stages of severe slugging in an S‐shaped riser 164
Figure 7‐12. Scale accumulation in choke (image taken from sandengen[7‐2] ) 165
Figure 7‐13. Erosion damage in a cage‐type choke [source unknown) 166
Figure 7‐14. CFD simulation of erosion in a production header 166
Figure 7‐15. a) Illustration or a corrosion reaction b) corrosion on a casing inside surface c) corrosion on tubing
167
Figure 7‐16. Wet gas flow in a horizontal flowline depicting top of line condensation 167
Figure 7‐17. Protective layer of FeCO3 formed on the metal surface b) inhibitors attached to the metal surface
168
Figure 7‐18. a) Wax crystals visible in a crude at WAT, b) WATs at different pressures in the phase diagram
168
Figure 7‐19. Crude oil not flowing once the pour point is reached 169
Figure 7‐20. a) wax plug retrieved topside (image taken from labes‐carrier et al[7‐3] ), b) evolution of the wax
thickness in a pipeline with time 169
Figure 7‐21. Flow schematic of a subsea production system with facilities for pigging and individual well testing
170
Figure 7‐22. a) oil (red) and water (White) originally separated, b) oil and water emulsion after vigorous stirring
in a blender. photos taken by hong[7‐4] 171
Figure 7‐23. Measured pressure drop in a horizontal pipe keeping the total flow rate constant and changing
water volume fraction, 𝒒𝒘/ 𝒒𝒘 𝒒𝒐 172
Figure 7‐24. oil‐water flow pattern map of water volume fraction versus mixture velocity for an upward pipe
inclination of 45°. figure adapted from rivera[7‐5] [7‐1] . 172
Figure 7‐25. Mixture viscosity behavior versus water volume fraction exhibited by the oil water mixture 173
Figure 8‐1. plot of factor F versus inlet fluid velocity that gives a change in specific kinetic energy equal to 0.981
m2/s2 178
Figure 8‐2. plot of pressure gradient versus inlet pressure that give a change in specific potential energy equal
to 0.981 m2/s2, for Vin=30 m/s 180
Figure 8‐3. plot of values of inlet velocity versus inlet pressure that give a change in specific kinetic energy
equal to 0.981 m2/s2, for several values of pressure gradient 181
Figure 8‐4. Sketch showing the side view of a subsea pipeline with inner forced convection, conduction in pipe
wall and insulation and free convection with water 183
Figure 8‐5. Sketch showing the side view of a wellbore with inner forced convection, conduction in tubing wall,
free convection in annulus, conduction in wall of production casing, conduction in cement layer between
production casing and intermediate casing, conduction in wall of intermediate casing, conduction in cement
layer between formation and intermediate casing 187
Figure 8‐6. Behavior of specific enthalpy of gas and oil vs. pressure for three temperatures 193
Figure 8‐7. Phase diagram of the hydrocarbon mixture used in Figure 8‐6 193

15
Figure 8‐8. Behavior of specific enthalpy of gas and oil vs. pressure for three temperatures 193
Figure 8‐9. Phase diagram of the hydrocarbon mixture used in Figure 8‐8 194
Figure 9‐1. Sample flow pattern map for vertical (90°) upward flow of gas and liquid 196
Figure 9‐2. Visualization of liquid loading transition criteria when the liquid is transported mainly as a film on
the pipe wall. a) represents an unloaded condition (annular flow) and b) represents a loaded condition (Churn‐
Annular). Taken from Van’t Westende[7‐3] 197
Figure 9‐3. Flow patterns along the tubing for the test well for several values of gas rate 198
Figure 9‐4. Flowing bottom‐hole pressure and percentage of wellbore volume occupied by the liquid versus
standard conditions gas rate 199
Figure 9‐5. Curves of critical gas rate versus water gas ratio for a wellhead pressure of 14.8 bara. 199
Figure 9‐6. Curves of critical gas rate versus water gas ratio for several values of wellhead pressure 200
Figure 9‐7. Curves of critical gas rate versus water gas ratio for several values of tubing inner diameter 200
Figure 9‐8. Sketch showing a configuration at well bottom with distributed inflow with downhole liquid
separation and back‐seepage from wellbore to the formation 201
Figure E‐1. Gas Processing from well to sales 226
Figure E‐2. Gas Processing from well to sales (including typical operating values) 227

16
 Field Performance M. Stanko

1. FIELD PERFORMANCE
The flow interaction between reservoir and production system defines the most important output of an oil
and gas asset: the production profile (the produced flow rates of oil or gas with time). The production profile
is one of the most important performance indicators of a field as it defines the revenue profile thus allowing
to compute the economic value of the asset.
The production profile is typically computed and predicted using analytical or numerical models (e.g.
simulators) that represent accurately the reservoir and production system. The fundamental idea is to produce
several times a “virtual field” testing different alternatives (e.g. production strategies, enhanced recovery
methods, etc.) to determine which one provides the best economic value. Once the best alternative is
determined, the production strategy is executed on the real asset.
This analysis is usually performed multiple times both during the field design phase and in the operational
phase. In the field design phase, the main goal is to compare different production and development strategies
and architectures. The numerical models are not yet fully defined and there are lot of uncertainties in the
input data. For an existing asset, it is usually used to foresee future problems, to evaluate the implementation
of Improved Oil Recovery (IOR) methods, drilling additional wells, among others. The numerical model is very
well defined and the historical production data has been used to reduce uncertainties in the models3 and
improve their predictability.
The two systems (reservoir and production system) are governed by different physical phenomena. However,
the field performance is defined by the interaction between them. When seen from the reservoir side, the
production system defines the back‐pressure acting on the sand‐face of the wells. When seen from the
production system side, the reservoir defines the amounts of fluids coming into the well and the formation
deliverability.

1.1. RESERVOIR
The reservoir is a heterogeneous porous media that contains oil, gas and water under pressure and where
wells have been drilled and completed. The wellbores are at a pressure lower than reservoir pressure which
causes the migration of fluid from the neighboring porous media to the wells. The flow deliverability of the
formation depends, among other things, on the pressure at the wellbore, the rock properties, the average
reservoir pressure, fluid properties, flow restrictions in the vicinity of the wellbore, extension and shape of the
drainage area. The deliverability of the reservoir will be typically reduced with time as fluids are drained from
it, the average pressure declines and the distribution and saturation of fluids in the reservoir changes.
A simplistic but useful analogy of a reservoir system is a tank with fluid under pressure inside. The well is a
small exit port with a restriction. The average reservoir pressure (i.e. the tank pressure, pR) drives fluid from
the tank to the wellbore (pwf, pressure at the exit). The restriction represents the pressure losses that are
generated when the fluid flows through the formation towards the well. When fluid is drained from the tank
(formation) the tank pressure (reservoir pressure) is reduced, thus reducing the flow rate that the tank can
deliver at a fixed wellbore pressure.

3
Reservoir models are typically history matched to production data. Production system models are typically tuned with
pressure, temperature and rate measurements along the production system.

17
 Field Performance M. Stanko

(A) (B)
FIGURE 1‐1. (A) TANK ANALOGY OF A (B) RESERVOIR SYSTEM

The main time scale of interest for field‐life studies is in the range of days‐weeks‐months‐years. Even though
there are also short transient events in the scale of hours, minutes and seconds (e.g. when the bottom‐hole
well conditions are changed suddenly, the well is closed for a period of time due to intervention, etc.) these
events are usually ignored. This is because they occur over a short period of time and thus they do not usually
affect the overall performance of the field (production, recovery factor, reservoir pressure, etc.).
The depletion performance of the reservoir is typically predicted using three approaches:
 Material balance
 Decline curve analysis
 Reservoir simulation
The second approach will not be discussed in this section.
In material balance, the reservoir is represented by a tank with oil, gas and water under pressure (Figure 1‐2).
Calculations are executed in a stepwise manner where the amount of oil or gas produced from the reservoir
is given as an input and the new saturation of fluids and pressure inside the tank are calculated by applying
conservation of mass in the tank. The producing gas oil ratio or water cut of the produced fluids can be
predicted using the change of the phase mobilities due to changes in phase saturation.

FIGURE 1‐2. GRAPHICAL DEPICTION OF THE MATERIAL BALANCE APPROACH

18
 Field Performance M. Stanko

A material balance model requires the oil (or gas) cumulative production as an input and thus cannot be used
to predict the production output of the reservoir with time. For that purpose, an additional model must be
provided to quantify the pressure drop between reservoir and a downstream condition (e.g. bottom‐hole
pressure). This model is often an Inflow Performance Relationship curve.
A reservoir simulator is used when it is important to consider the spatial (2D or 3D) variation of properties (e.g.
pressure, saturation) in the reservoir with time. The reservoir model consists of a numerical discretization of
the porous media where mass conservation is applied in every sub‐volume. The flow between cells is described
using an expression for pressure drop in porous media (e.g. Darcy’s Law). Pressure or rate boundary conditions
are applied on the cells where the wells are, and no flow conditions are typically applied at the outer edges of
the reservoir.
The model uses as input the initial distribution of pressure, porosities, permeabilities, fluid saturations, and it
computes the time evolution of pressure, oil, gas and water saturation. The simulation is controlled with both
a target rate and a minimum pressure at the well boundaries provided at each time step. The computation is
carried out in a stepwise manner, outputting results for pre‐specified time intervals.
During the solving process, the minimum pressure given is imposed on the well boundary. If the rate computed
is higher than the target rate specified, then the target rate is feasible. A series of iterations are then made
trying several pressure values until one value is found that gives exactly the target rate specified. On the other
hand, if the rate obtained is below the target rate, the target rate is not feasible, and the well boundary
condition is the minimum pressure.
Depending on the complexity of the field, in some cases it is possible to use only the reservoir simulator to
predict its performance and neglect the rest of the production system. For example, in a field where each well
is producing to its own separator close to the wellhead, including tubing pressure drop tables in a reservoir
model provides an exact approximation of the field performance.
In reservoir simulation, the grid does not typically capture the near‐wellbore region in detail. The well usually
traverses through several blocks and the block size is much bigger than the wellbore radius. In consequence,
an IPR‐like equation (often called well index or WI) must be used; this equation relates the formation oil, water
and gas with the pressure difference between the block where the well is placed and the wellbore pressure.

1.2. PRODUCTION SYSTEM (SURFACE NETWORK)

The production system is the assembly of wells, pipes, valves, pumps, meters that have the function of
transporting fluids from the reservoir to the processing facilities in a controlled manner. When the fluid travels
from the reservoir(s) (source) to the separator(s) (sink), it must overcome energy losses (e.g. pressure and
temperature drop) and sometimes “compete” with other fluids in transportation conduits.
In contrast with the reservoir, the field‐life analysis of a production system is performed assuming that changes
in reservoir deliverability are slow enough so that the system progresses continuously from one steady‐state
to another. Therefore, the analysis is usually performed at a given point in time, ignoring all past and all present
conditions and using only the current deliverability of the reservoir. Other possible quick transients such as
slugging, intermittent production, etc. are not part of the scope of a field‐life analysis.
In models of the production system, the well inflow at a particular time “t” is usually represented by an IPR
equation (Inflow performance relationship, see Figure 1‐3 for some examples). The IPR is typically a smooth,
monotonic, downwards curve that provides the bottom‐hole pressure that must be applied at the sand face
to deliver a specific standard condition flow rate. This approach is usually a good approximation to reservoir
deliverability.

19
 Field Performance M. Stanko

IPR for undersaturated oil IPR for saturated oil/gas

FIGURE 1‐3. IPR CURVE

The IPR curves come typically from recent well tests, by using analytical equations together with limited field
data, or generated by a reservoir simulator.
The produced rates, pressures and temperatures at time “t” are calculated by performing a flow equilibrium
calculation in the production system. This involves solving simultaneously mass, momentum and energy
conservation equations for all elements in the system (conduits, flowlines, pipelines, valves, pumps, etc).
Pipelines are typically discretized in segments. The boundary conditions upstream are the IPRs, (i.e. the wells’
inflows) and downstream the pressure(s) of the separator(s).
When there is adjustable equipment in the production system (e.g. adjustable chokes, pumps, gas lift injection)
there is usually a variety of “feasible” equilibrium rates that the system can produce. For example, in a system
with a choked well, the rate of the well can vary depending if the choke is fully open, fully closed or something
in between. If the well has an electric submersible pump (ESP) then a variety of operational rates can be
achieved by changing the pump rotational speed.

1.3. COUPLING RESERVOIR MODELS AND MODELS OF THE PRODUCTION SYSTEM

As mentioned earlier, the production profile of the field should be computed considering the interaction
between the reservoir and production system. Figure 1‐4 shows a possible way to couple a material balance
(MB) model of the reservoir with a model of the production system to obtain the production profile of the
field.

20
 Field Performance M. Stanko

Material balance model Production system Model

Initial timestep Results, pRi, Compute

“t1” Transfer to
Sgi,Soi,Swi wells’ IPRs

Settings of
Run the adjustable
model elements (if
any)
Compute
Time Results:
Cumulative
step t production Transfer to
Well rates
and
Qp*
pressures

Solve the
MB
equations

* Assuming that the rate

Timestep “ti+1” Results
remains constant during
the production interval

FIGURE 1‐4. TIME STEP OF AN EXPLICIT COUPLING SCHEME BETWEEN A MATERIAL BALANCE MODEL AND A MODEL OF THE
PRODUCTION SYSTEM TO PREDICT PRODUCTION PROFILE

Since the model of the production system is steady‐state, the changes associated with reservoir depletion are
introduced by modifying the IPRs in every time step (based on the output from the reservoir model). In this
particular case, the IPR is recalculated in every time step using the reservoir pressure and the mobility of the
oil and gas phases (calculating relative permeability from the saturation).
When using a reservoir simulation model, the IPR curves are often generated by the reservoir simulator and
transferred to the network model. An example of this methodology is shown in Figure 1‐5.

Reservoir Model Network Model

Reservoir Calculate
timestep “ti” well IPR

IPRs of
Input
each Transfer to
data
well

Settings of
Run the adjustable
model elements (if
any)

Well
Well Transfer to rates and
controls
pressures

Execute
time step

Reservoir
timestep “ti+1” Results

FIGURE 1‐5. EXPLICIT COUPLING BETWEEN A RESERVOIR SIMULATION AND A MODEL OF THE PRODUCTION SYSTEM

There are multiple approaches to couple reservoir and production system models. The two approaches
discussed before are explicit because, for a given time step, rates are calculated only once in the network
model and then imposed in the reservoir model. The rates are then assumed to remain constant during the
time interval specified. However, this is seldom the case because of the reduction of reservoir pressure when
fluids are produced. A workaround frequently applied to reduce this inaccuracy is to reduce the length of the

21
 Field Performance M. Stanko

time step. Explicit coupling strategies sometimes cause instabilities in the solution (oscillating production rates
with time). The reduction of time step length often eliminates this problem.
Explicit coupling strategies are suitable when the models of reservoir and production system are available in
two separate computational routines (often black box commercial packages) and are maintained and used
separately (e.g. by different departments within the company). An explicit coupling minimizes the required
transfer of information between models in every time step.
Other coupling approaches and a detailed discussion about coupling models or the reservoir and the
production system are discussed in detail by Barroux[1‐1].

1.4. PRODUCTION POTENTIAL

For a particular time “t” there will be either a unique rate that the field can produce (if there are no adjustable
elements in the system of they have a fixed setting) or a maximum rate that the field can produce (if there are
adjustable elements). We will refer to this unique or maximum rate that the field can produce at a given point
in time as: “production potential”.
For example, if the well has an adjustable choke the maximum rate is most likely achieved when the choke is
fully open. If the well has an electric submersible pump (ESP) then the maximum rate is probably achieved
when the pump rotational speed is highest. If adjustable elements are present in the system, it is usually
possible to produce any rate lower than the maximum rate by regulating such elements.
The production potential is different from the “well potential” variable printed in every time step by the
reservoir simulator. The well potential is the producing rate obtained when the minimum bottom‐hole
pressure is applied on the well boundary.
To illustrate how these two concepts are different, consider a single well system in which wellhead pressure
is kept constant. The well potential of the reservoir simulator is estimated using a constant bottom‐hole
pressure as shown in Figure 1‐6.b, only taking into account the reservoir deliverability (inflow performance
relationship). The production potential is calculated by performing a hydraulic equilibrium calculation at the
bottom‐hole intersecting the IPR and tubing performance relationship (TPR) shown in Figure 1‐6a. These two
values will be equal only when the minimum bottom‐hole pressure specified equals the equilibrium bottom‐
hole pressure (in the fig. when pR = pR3). For the other IPRs however, the production potential is over‐
predicted.

A) B)

FIGURE 1‐6. (B) WELL POTENTIAL CALCULATION VS. (A) PRODUCTION POTENTIAL CALCULATION

As time progresses, the production potential of the production system will also change, mainly due to two
types of changes: changes in the inflow (IPR) and changes to the production system.

22
 Field Performance M. Stanko

A strategy commonly used in reservoir management when producing the field in plateau mode is to allocate
production to individual wells using their potential. At a given time, the well and field potential are calculated,
and well split factor are computed by dividing the individual well production potential by the field production
potential. Then, the rate to be produced by each well is calculated by multiplying the field plateau rate by the
individual well split factor.
In a producing field, the reservoir deliverability follows a trend with depletion similar to reservoir pressure, i.e.
is reduced with time. This is not only due to reservoir pressure decline, but for example, in an oil well, an
increase of the well’s producing GOR and WC will reduce the oil productivity as well. Changes in reservoir
deliverability affect all components of the production system downstream the reservoir, for example if the
well producing gas oil ratio (GOR) changes, then pressure losses will change in all downstream conduits.
Some examples of changes to the production system are man‐made changes in the pipeline diameter,
lowering separator pressure, modification of choke opening, changes in well completion, installation of
artificial lift, well stimulation or fracking, etc. Other changes are reduction of the conduits’ cross section due
to scale deposition, wax deposition, etc. When the modification is abrupt and occurs at one point in time, the
production potential will display a discontinuity at the particular cumulative production where the change is
introduced (as shown in Figure 1‐7).

FIGURE 1‐7. PRODUCTION POTENTIAL BEHAVIOR VS. TIME WHEN A PRODUCTION ENHANCEMENT MODIFICATION IS PERFORMED
IN THE SYSTEM

The decrease in reservoir deliverability causes a decrease in production potential with time. Changes in the
production system can increase or decrease the production potential with time, depending on the type of
change as explained in the previous paragraph.

1.5. PRODUCTION SCHEDULING

There are two main types of production offtake in a field: period with fixed production rate (plateau mode) or
declining production (decline mode). In plateau mode, as the name indicates, the field or well is produced at
a constant rate for a given period (lower than the production potential). However, as the production potential
is typically reduced with time, there comes a time when the field rate is the same as the production potential.
After that moment, the field will not be able to sustain the plateau rate and its production starts to decline
(e.g. following the production potential curve). This is shown in Figure 1‐8.

23
 Field Performance M. Stanko

FIGURE 1‐8. PLATEAU PRODUCTION MODE

This production mode is typically employed for standalone field developments with dedicated processing
facilities or when there are contractual production obligations (e.g. gas contracts). This is usually the outtake
strategy that yields the best economic value for the project. Producing more at an early stage (like in decline
mode) increases revenue but the CAPEX investment becomes excessive due to the increased size of the
processing facilities and offshore structure.
In decline mode, as the name indicates, production rates typically decline with time (as shown in Figure 1‐9).
In principle, the objective is to produce as much as possible as early as possible (i.e. always produce at the
production potential of the system). However, the production rates might be sometimes lower than the
production potential but follow a similar decline with time. This may occur for example when there are
additional operational constraints that impede reaching the production potential, e.g. maximum flow rate to
avoid sand production, gas coning, water coning, maximum drawdown in the formation.
It can also occur when the adjustable equipment is operated at a constant setting or there are non‐trivial
settings (unknown to the field operator) of the adjustable equipment that yield maximum production (e.g. a
particular gas‐lift rate that yields optimum oil production).

FIGURE 1‐9. PRODUCTION PROFILE OBTAINED WHEN OPERATING IN DECLINE MODE

This production mode is employed typically for satellite fields that will use the spare capacity of the processing
facilities of a neighboring mature field.
Figure 1‐10 shows the production profile of the single well production system producing in plateau mode and
producing at the production potential from the beginning. Both systems are being produced until the same
ultimate amount of gas or oil is recovered (QPU).
Fixing a constant plateau rate causes that the time to abandonment is considerably prolonged when compared
to the open choke case. However, the total amount of oil or gas recovered is the same (i.e. blue and violet
areas are the same). A lower plateau rate will give an even longer time to abandonment.

24
 Field Performance M. Stanko

FIGURE 1‐10. PRODUCTION RATE BEHAVIOR VS CUMULATIVE PRODUCTION FOR OPEN CHOKE AND CONSTANT RATE

1.6. RELATIONSHIP BETWEEN PRODUCTION POTENTIAL AND CUMULATIVE PRODUCTION

As mentioned earlier, the production potential partly depends on the deliverability of the formation. In reality,
the reservoir deliverability (IPR) does not depend on time but mainly on the amount of fluid that has been
withdrawn from the reservoir since the initial condition to time t:

EQ. 1‐1
𝑄 𝑞 𝑡 ∙ 𝑑𝑡

QP represents amounts of oil or gas and it is called cumulative production.

Figure 1‐11 shows the Well IPR curve (flowing bottom‐hole pressure vs flow rate) for 3 increasing values of
cumulative production. Note that the IPR of the formation is reduced if more fluids have been produced from
the reservoir.

FIGURE 1‐11. CHANGES OF IPR WITH CUMULATIVE PRODUCTION

This implies that the production potential at a given point in time is mainly dependent on the how much fluid
has been produced up to that point in time.
If a method of estimating well IPR vs. cumulative production is available, the computation of the production
potential curve using numerical models is straightforward.
Example 1: Consider a production system consisting of an undersaturated oil reservoir with an underlying
aquifer with volume Va and a number of identical wells Nw. The pressure will decline according to Eq. 1‐2:
EQ. 1‐2
𝑝 𝑝 𝑁 ∙𝐴

With A being:

25
 Field Performance M. Stanko

𝐵
𝐴 EQ. 1‐3
𝑐 ∙𝑆 𝑐
𝑁∙𝐵 , ∙ 𝑐 𝑉 ∙𝜙 ∙ 𝑐 𝑐 ∙𝐵
𝑆
The rate of a single well can be expressed as:
EQ. 1‐4
𝑞 , 𝐽∙ 𝑝 𝑝

If the well has some sort of artificial lift method installed such as an electric submersible pump, usually the
maximum well rate will be achieved when the flowing bottom‐hole pressure is lowered to a minimum value:
EQ. 1‐5
𝑞 , 𝐽∙ 𝑝 𝑝 ,

Considering the number of wells and that the effect of the flow commingling in the surface network does not
affect the individual well performance, then the field maximum rate can be expressed as:
EQ. 1‐6
𝑞 , 𝑁 ∙𝐽∙ 𝑝 𝑝 ,

Substituting Eq. 1‐2 in Eq. 1‐6:

EQ. 1‐7
𝑞 , 𝑁 ∙𝐽∙ 𝑝 𝑁 ∙𝐴 𝑝 ,

Expanding the terms and grouping:

EQ. 1‐8
𝑞 , 𝑁 ∙𝐽∙𝑁 ∙𝐴 𝑁 ∙𝐽∙ 𝑝 𝑝 ,

Renaming terms:
EQ. 1‐9
𝑞 , 𝑞
EQ. 1‐10
𝑞 𝑁 ∙𝐽∙ 𝑝 𝑝 ,

EQ. 1‐11
𝑚 𝐴∙𝑁 ∙𝐽

Finally, the production potential is given by the following expression:

𝑞 𝑚∙𝑁 𝑞 EQ. 1‐12

The production potential of the system will follow the smooth, downwards and continuous trend shown in
Figure 1‐12.

FIGURE 1‐12. PRODUCTION RATE BEHAVIOR VS CUMULATIVE PRODUCTION FOR OPEN CHOKE SHOWING THE REGION OF FEASIBLE
RATES

26
 Field Performance M. Stanko

Example 2: Consider a production system where there are 𝑁 identical wells producing from a common dry
gas reservoir, each one with their own separator and flowline. The dry gas tank material balance equation is:

𝑍 ∙𝑝 𝐺 EQ. 1‐13
𝑝 ∙ 1
𝑍 𝐺

The production of a single well can be expressed with the low‐pressure backpressure equation as a function
of the field rate (𝑞 ), the total number of wells (𝑁 ):

𝑞 EQ. 1‐14
𝐶∙ 𝑝 𝑝
𝑁

The dry gas tubing equation is:

.
𝑞 𝑝 EQ. 1‐15
𝐶 ∙ 𝑝
𝑁 𝑒

Finally, the flowline equation (assuming horizontal flowline):

𝑞 . EQ. 1‐16
𝐶 ∙ 𝑝 𝑝
𝑁

To compound everything in one equation, one follows the procedure:

1. Clearing 𝑝 from Eq. 1‐16, then substituting in Eq. 1‐15.
2. Clear 𝑝 from the equation found in step 1 and substitute in Eq. 1‐14.
3. Clear 𝑝 from the equation found in step 2 and substitute in Eq. 1‐13. This gives:

.
𝑞 𝑞 𝑞 𝑍 ∙𝑝 𝐺 EQ. 1‐17
𝑒 ∙ 𝑒 ∙ 𝑒 ∙𝑝 ∙ 1
𝑁 ∙𝐶 𝑁 ∙𝐶 𝑁 ∙𝐶 𝑍 𝐺

Eq. 1‐17 is plotted in Figure 1‐13 using the following input:

 Number of wells, 𝑁 5, 10, 15

 Backpressure coefficient, 𝐶 1000 𝑆𝑚 ⁄𝑏𝑎𝑟 ∙
 Backpressure exponent, 𝑛 1
 Tubing elevation coefficient, 𝑠 0.155
 Tubing coefficient, 𝐶 4.03 ∙ 10 𝑆𝑚 ⁄𝑏𝑎𝑟
 Flowline coefficient, 𝐶 2.83 ∙ 10 𝑆𝑚 ⁄𝑏𝑎𝑟
 Separator pressure, 𝑝 30 𝑏𝑎𝑟𝑎
 Initial reservoir pressure, 𝑝 276 𝑏𝑎𝑟𝑎
 Gas deviation factor calculated with the correlation of Hall and Yarborough, T = 92 oC and gas specific
gravity 0.5.
 Initial gas in place 𝐺 2.7 ∙ 10 𝑆𝑚

27
 Field Performance M. Stanko

FIGURE 1‐13. FIELD PRODUCTION POTENTIAL VS CUMULATIVE PRODUCTION FOR DRY GAS RESERVOIR WITH STANDALONE WELLS

When more wells are used in the system, the production potential is higher. The effect is proportional, due to
the fact that well are standalone, and adding a new well does not interfere or affect the performance of other
wells.
Figure 1‐14 shows the dimensionless production potential of the field versus recovery factor. The
dimensionless production potential has been found by dividing each field production potential curve by its
maximum production potential (i.e. the production potential at Gp = 0 Sm3) and the cumulative production by
the initial gas in place. Surprisingly, the curves for all number of wells fall on top of each other.

FIGURE 1‐14. DIMENSIONLESS FIELD PRODUCTION POTENTIAL VS RECOVERY FACTOR FOR DRY GAS RESERVOIR WITH STANDALONE
WELLS

The dimensionless production potential curve in Figure 1‐14 remains unchanged if the amount of gas in place
is increased or decreased. Figure 1‐15 shows the dimensionless production potential of the field estimated

28
 Field Performance M. Stanko

with variations of ±50% on the backpressure coefficient, tubing coefficient, flowline coefficient and varying
separator pressure. These variations cause modest changes in the curve.

FIGURE 1‐15. DIMENSIONLESS FIELD PRODUCTION POTENTIAL VS RECOVERY FACTOR FOR DRY GAS RESERVOIR WITH STANDALONE
WELLS, SENSITIVITY STUDY ON SYSTEM PROPERTIES

Example 3: Consider a production system where there are 𝑁 identical wells producing from a common dry
gas reservoir, but they are grouped in 𝑁 templates (where each template has a number of 𝑁 , wells). There
are identical flowlines from each template to a common junction, and one long pipeline from the junction to
the separator. The dry gas tank material balance equation is:

𝑍 ∙𝑝 𝐺 EQ. 1‐18
𝑝 ∙ 1
𝑍 𝐺

The production of a single well can be expressed with the field rate (𝑞 ), the total number of wells per template
(𝑁 , ), the total number of wells and the low‐pressure backpressure equation:

𝑞 EQ. 1‐19
𝐶∙ 𝑝 𝑝
𝑁 , ∙𝑁

The dry gas tubing equation is:

.
𝑞 𝑝 EQ. 1‐20
𝐶 ∙ 𝑝
𝑁 , ∙𝑁 𝑒

The flowline equation (assuming horizontal flowline):

𝑞 . EQ. 1‐21
𝐶 ∙ 𝑝 𝑝
𝑁

The pipeline equation (assuming horizontal pipeline):

29
 Field Performance M. Stanko

. EQ. 1‐22
𝑞 𝐶 ∙ 𝑝 𝑝

To compound everything in one equation, one follows the procedure:

1. Clearing 𝑝 from Eq. 1‐22, then substituting in Eq. 1‐21.
2. Clear 𝑝 from the equation found in step 1 and substitute in Eq. 1‐20.
3. Clear 𝑝 from the equation found in step 2 and substitute in Eq. 1‐19.
4. Clear 𝑝 from the equation found in step 3 and substitute in Eq. 1‐18. This gives:

𝑞 𝑞 𝑞
𝑒 ∙ 𝑒 ∙ 𝑒 ∙𝑝 𝑒
𝑁 , ∙𝑁 ∙𝐶 𝑁 , ∙𝑁 ∙𝐶 𝐶
EQ. 1‐23
.
𝑞 𝑍 ∙𝑝 𝐺
∙ ∙ 1
𝐶 ∙𝑁 𝑍 𝐺

To evaluate the effect of the gathering system on the production potential curve, the expressions for
standalone wells (Eq. 1‐17), network4 wells (Eq. 1‐23) and considering IPR only (with fixed bottom‐hole
pressure of 120 bara) are plotted in Figure 1‐16. The network doesn’t affect significantly the dimensionless
field production potential curve when compared to the standalone case but excluding the system downstream
the well bottom‐hole does.

FIGURE 1‐16. DIMENSIONLESS FIELD PRODUCTION POTENTIAL VS RECOVERY FACTOR FOR DRY GAS RESERVOIR WITH STANDALONE
WELLS, NETWORK WELLS AND CONSIDERING IPR ONLY.

Example 4: Figure 1‐17 shows the dimensionless field production potential vs. recovery factor for several cases.
All cases had a surface gathering network transporting production to the processing facilities coupled with the
reservoir model.

4
Using flowline coefficient Cfl = 2.83 105 Sm3/bar and pipeline coefficient Cpl = 2.75 105 Sm3/bar

30
 Field Performance M. Stanko

FIGURE 1‐17. DIMENSIONLESS FIELD PRODUCTION POTENTIAL VS RECOVERY FACTOR FOR SEVERAL PRODUCTION SYSTEMS.

More details about the cases, modeling assumptions and results can be found in Stanko [1‐4].
A general continuous equation to reproduce most of these curves is a fifth order polynomial:
EQ. 1‐24
𝑞 𝑎 ∙ 𝑅 𝑎 ∙ 𝑅 𝑎 ∙ 𝑅 𝑎 ∙ 𝑅 𝑎 ∙𝑅 1

If, on the other hand, the curves have some discontinuity or do not fit well to a continuous function, then a
piece‐wise linear approximation can be a more suitable alternative:

𝑞 , 𝑞 ,
⎧ ∙ 𝑅 𝑅 , 𝑞 , 𝑖𝑓 𝑅 , 𝑅 𝑅 , ⎫
𝑅 , 𝑅 ,
⎪ ⎪
⎪ 𝑞 , 𝑞 , ⎪
⎪ ∙ 𝑅 𝑅 , 𝑞 , 𝑖𝑓 𝑅 , 𝑅 𝑅 , ⎪
𝑅 , 𝑅 ,
𝑞 EQ. 1‐25
𝑞 , 𝑞 ,
⎨ ∙ 𝑅 𝑅 , 𝑞 , 𝑖𝑓 𝑅 , 𝑅 𝑅 ,
⎬
⎪ 𝑅 , 𝑅 , ⎪
⎪𝑞 … ⎪
⎪ , 𝑞 , ⎪
∙ 𝑅 𝑅 , 𝑞 , 𝑖𝑓 𝑅 , 𝑅 𝑅 ,
⎩ 𝑅 , 𝑅 , ⎭

Where 𝑞 , ,𝑅 , are data point pairs, and there is a total of “N” point pairs.

The curves of production potential versus cumulative production of a production system can be computed
from these curves by multiplying 𝑞 by the maximum production of the system at recovery factor equal zero
(𝑞𝑝𝑝𝑜,𝑓𝑖𝑒𝑙𝑑 ), and by substituting recovery factor by Qp/Q. To update the curves of production potential in time,
if, for example, at some point in time, a new well is drilled, then the value of 𝑞 , must be updated
accordingly.

1.7. PRODUCTION SCHEDULING AND PLANNING USING PRODUCTION POTENTIAL CURVE

The production potential curve versus cumulative production can be used to perform production scheduling
and planning without recurring to perform coupled runs of reservoir and production models. This is because

31
 Field Performance M. Stanko

by expressing the production potential as a function of cumulative production, the time dependency has been
removed. To ilustrate this, three examples will be presented and discussed next.
Example 1: Undersaturated oil reservoir with an underlying aquifer with volume Va and a number of identical
wells Nw.
Let us assume the field will be produced at a plateau rate (𝑞 , ) initially and then it will enter in decline. The
plateau will end when the field production potential becomes equal to the field plateau rate (𝑞 , )

EQ. 1‐26
𝑞 , 𝑚 ∙ 𝑁∗ 𝑞

Then, the plateau duration can be calculated with the cumulative production 𝑁 ∗

𝑁∗ 𝑞 , 𝑞 𝑞 1 EQ. 1‐27
𝑡 1 ∙
𝑞 , 𝑞 , ∙𝑚 𝑞 , 𝑚
After the plateau, the field will produce at potential:
EQ. 1‐28
𝑞 𝑞 , 𝑚∙𝑁 𝑞

Expanding the definition of cumulative production:

EQ. 1‐29
𝑞 𝑚 ∙ 𝑁∗ 𝑞 ∙ 𝑑𝑡 𝑞

Substituting the definition of 𝑁 ∗

𝑞 𝑞 EQ. 1‐30
𝑞 𝑚∙ 𝑞 ∙ 𝑑𝑡 𝑚∙ 𝑞
𝑚

Simplifying:

𝑞 𝑚∙ 𝑞 ∙ 𝑑𝑡 𝑞 EQ. 1‐31
_

A solution to this equation is:

∙ EQ. 1‐32
𝑞 𝑞 , ∙𝑒

Therefore, if the production potential displays a linear behavior with respect to cumulative production, the
production profile post‐plateau has an exponential behavior with time. The coefficient of the exponential
function, that dictates the rate decline depends both on the decline characteristics of the reservoir (A), the
flow “resistance” in the formation (and in principle, in the tubing and surface flowlines) and the number of
wells. If the number of wells is increased, the decline will become more pronounced.
The field production profile is given by the following equations:

𝑓𝑜𝑟 𝑡 𝑡 𝑞 𝑞 ,
EQ. 1‐33

𝑓𝑜𝑟 𝑡 𝑡 𝑞 𝑞 ∙𝑒 ∙ EQ. 1‐34

,

32
 Field Performance M. Stanko

Example 2: Plateau mode production

Consider the production strategy proposed in Figure 1‐18a. The production potential curve has been divided
in three parts that will be produced at constant rate. The production rates are feasible because they fall below
the production potential line. Figure 1‐18b. shows the production profile calculated from Figure 1‐18a. As the
reservoir is produced with constant rate periods, it is simple to estimate the duration of each period by dividing
the cumulative production of the period by the period rate.

(A) (B)
FIGURE 1‐18. PLATEAU MODE PRODUCTION

For a few simple cases (e.g. dry gas, undersaturated oil) an analytical expression of the production potential
𝑞 can be found. However, in general, for the majority of cases (e.g. saturated oil, gas condensate) this is not
possible. The production potential must then be calculated by running a simulation of coupled reservoir and
production models at maximum rate and record the field rate and the cumulative production Qp. This process
yields a collection of points.
For cases where the production potential is not linear, it is usually not practical to solve analytically for the
plateau duration and post‐plateau field rate as presented in the previous examples. If an analytical expression
is available, plateau duration can be estimated by substituting the desired plateau rate and solve the equation
(usually with a root solving method) for the cumulative production at plateau end 𝑄∗ . If a collection of points
is available, 𝑄 ∗ can be found by interpolating on the table. With 𝑄∗ and plateau rate, one can then calculate
plateau duration.
The post‐plateau field rate can be estimated by dividing the post‐plateau period in a series of discrete time
steps and expressing the cumulative production at time ti using the trapezoidal rule for numerical integration:

𝑄 𝑡 0.5 ∙ 𝑞 𝑡 𝑞 𝑡 ∙ 𝑡 𝑡 𝑄 𝑡 EQ. 1‐35

All rates in the post‐plateau period should fall on the production potential curve, i.e.
EQ. 1‐36
𝑞 𝑡 𝑓 𝑄 𝑡

Eq. 1‐35 and Eq. 1‐36 must be solved simultaneously for each time step ti and departing from plateau end. If
the production potential is available as a collection of points, Eq. 1‐36 means interpolation.
Example 3: Production potential of a system with two standalone wells
Consider a field with two (2) standalone wells, and that the production potential of each well can be expressed
as a function of the cumulative production of each individual well:

𝑞 𝑓 𝑄 EQ. 1‐37

33
 Field Performance M. Stanko

In this case the production profile can be computed separately for each well from the production potential
curve and then add them up to obtain the field production profile. Note that the field production potential for
a given field cumulative production is not unique. This is because there are different ways to achieve the same
field cumulative production (e.g. in a two well system, produce more from well 1 than 2, produce equal, or
produce more from well 2 than 1).
As an example, consider the production system with 2 standalone wells shown in Figure 1‐19a. The production
potential of each well is presented in Figure 1‐19b. Wells will be produced at constant rate initially, with
plateau rates qP1 and qP2 and, when the plateau rate is no longer feasible, they will be produced at the
production potential.

(A) (B)
FIGURE 1‐19. EXAMPLE CASE: 2 STANDALONE WELLS

The plateau duration of each well can be very easily calculated by intersecting the individual plateau rate with
the production potential curve of each well. This yields a plateau duration of tp1 = QP1/qP1, for well 1 and tp2 =
QP2/qP2 for well 2. After the plateau ends, the production profile of each well follows the potential.
A typical reservoir management problem consists of how to define well rates to maximize field plateau
duration when a fixed field rate is desired. If individual well plateau rates are to be kept constant, this can be
achieved by finding the plateau rates for which the plateau end occurs at the same time. If the production
potential curves are straight lines the following procedure is suitable:
The production potential curve for well 1:
EQ. 1‐38
𝑞 𝑚 ∙𝑄 𝑞

The cumulative production at which the production potential (qpp1) is equal to the plateau rate (qp1), i.e. QPp1,
is:
𝑞 𝑞 EQ. 1‐39
𝑄
𝑚
Similarly, for well 2:
𝑞 𝑞 EQ. 1‐40
𝑄
𝑚
Then the plateau duration has to be the same for both wells:

𝑄 𝑄 EQ. 1‐41
𝑡 ;𝑡
𝑞 𝑞

34
 Field Performance M. Stanko

Substituting Eq. 1‐39 and Eq. 1‐40 in Eq. 1‐41:

𝑞 𝑞 𝑞 𝑞 EQ. 1‐42
𝑚 ∙𝑞 𝑚 ∙𝑞
𝑞 𝑚 𝑞 EQ. 1‐43
1 ∙ 1
𝑞 𝑚 𝑞

Eq. 1‐43 has two unknowns, therefore one more equation is needed. Clearing qp2 from the expression of the
total plateau rate:

𝑞 𝑞 𝑞 EQ. 1‐44

Substituting Eq. 1‐44 in Eq. 1‐43 yields:

𝑞 ∙ 𝑚 𝑚 𝑞 ∙ 𝑞 ∙𝑚 𝑞 ∙𝑚 𝑞 ∙𝑚 𝑞 ∙𝑚 𝑞 ∙𝑚 ∙𝑞 EQ. 1‐45
0
Eq. 1‐45 can be solved with the quadratic formula to find qp1:

𝑎 𝑚 𝑚 EQ. 1‐46

𝑏 𝑞 ∙𝑚 𝑞 ∙𝑚 𝑞 ∙𝑚 𝑞 ∙𝑚 EQ. 1‐47

𝑐 𝑞 ∙𝑚 ∙𝑞 EQ. 1‐48

𝑏 √𝑏 4∙𝑎∙𝑐 EQ. 1‐49

𝑞
2∙𝑎
Note that the main constraints used to solve this problem were that both wells must produce in plateau mode
with a constant rate and then will enter in decline at the same time. However, there are infinite alternatives
to produce the field at plateau rate as shown in Figure 1‐20 and each option will yield a different field plateau
duration.

35
 Field Performance M. Stanko

q_well_1 q_well_1

q_well_2 q_well_2

qpp_well_1 qpp_well_1

flow rate, [Sm^3/d]

flow rate, [Sm^3/d]

qpp_well_2 qpp_well_2

q_field q_field

0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
time, t, [t] time, t, [t]

(A) (B)

q_well_1 q_well_1

q_well_2 q_well_2

qpp_well_1 qpp_well_1
flow rate, [Sm^3/d]

flow rate, [Sm^3/d]

qpp_well_2 qpp_well_2

q_field q_field

0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
time, t, [t] time, t, [t]
5
(C) (D)
FIGURE 1‐20. 4 DIFFERENT ALTERNATIVES TO PRODUCE THE TWO WELLS SYSTEM IN PLATEAU MODE

1.8. APPLICABILITY OF THE PRODUCTION POTENTIAL CONCEPT IN REAL FIELDS AND MULTI‐WELL
PRODUCTION SYSTEMS

The production potential concept is valid only when the reservoir (or well) producing GOR, WC, reservoir
pressure and IPR can be safely predicted as a function of its cumulative production and the transient period
(infinite acting) is short.
This concept can be used to design and predict the production profile of the field. For example, in early stages
of field planning an assumption typically made is that all wells are identical. In consequence, the production
potential of the field is just the multiplication of the number of wells time the production potential of a single
well. This approach is often used to estimate roughly the number of wells that are required to produce the
field rate for a desired time period. An example of this method is presented and discussed thoroughly by Van
Dam [1‐5].
For more complex cases, e.g. reservoirs that cannot be modeled with a material balance approach, an iterative
approach is often used (without using the production potential method) where field models are run multiple
times with different production rates and their results compared.
More details about the limitations and applications of production potential curves can be found in Stanko [1‐4].

36
Field Performance M. Stanko

REFERENCES
[1‐1] Barroux, C., Duchet‐Suchaux, P., Samier, P. & Nabil, R. (2000). Linking Reservoir and Surface
Simulators: How to improve the Coupled Solutions. SPE‐65159. European Petroleum Conference.
Paris: Society of Petroleum Engineers.
[1‐2] Golan, M.; Whitson, C. H. (1986). Well Performance. Second Edition. Prentice‐Hall Inc. Englewood
Cliffs, New Jersey.
[1‐3] Nind, T. (1964). Principles of Oil Well Production. McGraw‐Hill.
[1‐4] Stanko, M. (2021). Observations on and use of curves of current dimensionless potential versus
recovery factor calculated from models of hydrocarbon production systems. Journal of Petroleum
Science and Engineering. Volume 196. 108014. ISSN 0920‐4105,
https://doi.org/10.1016/j.petrol.2020.108014.
[1‐5] Van Dam, J. (1986). Planning of Optimum Production from a Natural Gas Field. Journal of the
Institute of Petroleum 54 (521).

37
 Flow Performance in Production Systems M. Stanko

2. FLOW PERFORMANCE IN PRODUCTION SYSTEMS

The production system is the assembly of wells, pipes, valves, pumps, meters that have the function of
transporting fluids from the reservoir to the processing facilities in a controlled manner. Formally, the
processing facilities5 should also be considered as part of the production system but they are excluded from
the current discussion. This is because the primary separator pressure is usually kept constant (e.g. with a
control system as shown in Figure 2‐1) which decouples (in terms of flow and pressure dependence) the
system upstream and downstream the separator.

FIGURE 2‐1. SIMPLIFIED LEVEL AND PRESSURE CONTROL SYSTEM IN A SEPARATOR

The layout and characteristics of the production system might vary significantly depending on the reservoir
characteristics, its geographical location (offshore, onshore, remote access), the field development concept,
the existence of neighboring fields, among others. However, it is possible to define two clear configurations:
standalone wells (e.g. gas wells in domestic US) where each well is producing through their own pipeline to a
separator (as in Figure 2‐2a) or surface networks where well production is gathered by pipelines that
transverse the field and converge in the main production facilities (as in Figure 2‐2b).

(A) (B)
FIGURE 2‐2. LAYOUT OF TWO PRODUCTION SYSTEMS

The surface connectivity between wells defines, to a great extension, the degree of flow interference between
them (i.e. how the operating conditions in one well affect others).
When the fluid travels from the reservoir(s) (source) to the separator(s) (sink), it has to overcome energy losses
(e.g. pressure drop) and sometimes “compete” with other fluids in the transportation pipes. A flow equilibrium

5
An overview of typical parts of offshore oil and gas processing facilities is presented in Appendix E.

38
 Flow Performance in Production Systems M. Stanko

state is reached where the producing rates, pressures and temperatures of the system are a product of a
balance between the capacity of each source and the existing energy losses/additions.
Numerical models are often used to understand and estimate the flow equilibrium state of production
systems. The numerical model of a production system is usually a steady state representation that comprises
from the well bottom‐holes (source nodes) to the first stage separator(s) (sink nodes). The main purpose of
this model is to compute the rates from each well and the pressure and temperature distribution in the
production system.
The well inflow is typically represented by an IPR equation (Inflow performance relationship) that provides the
bottom‐hole pressure that has to be applied at the sand face to deliver a specific standard condition rate (see
Figure 2‐3). The IPR describes the reservoir deliverability for a given depletion state and assuming that a
pseudo‐steady state has been reached in the reservoir. Please note that the same well might be producing
from different reservoir regions or have several laterals, so several IPRs might be required for the same well.

IPR for undersaturated oil IPR for saturated oil/gas

FIGURE 2‐3. IPR CURVE

Flow in tubular conduits such as tubing, casing and pipelines is represented with equations that predict the
temperature and pressure drops6. Usually, these equations use constant fluid properties, so a length
discretization and a step‐wise calculation has to be performed to capture fluid behavior. The separator is
represented by a constant pressure value. Other elements, such as restrictions, chokes, valves, boosters, etc.
have their own particular equations to predict pressure and temperature change according to the energy
introduced or removed from the fluid.
These equations are usually derived by applying mass, momentum and energy conservation equations to the
element or interest. The equations are further simplified to reduce the number of unknowns by introducing
relationships between variables, empirical correlations, etc.
This set of equations that constitute the numerical model of the production system is solved simultaneously
in an iterative manner (e.g. using a Newton method). They have to be solved simultaneously because the
upstream or downstream conditions of one element are usually the downstream or upstream conditions of
another element. The solving process is usually referred to as computing the flow equilibrium of the
production system and usually consists on assuming and varying well rate(s) to minimize a pressure residual
using a Newton method.
In a single well‐pipeline‐separator system (as in Figure 2‐2a) the procedure might be as follows:

 Assume well rate

6
As an example, please refer to appendix A for the full development of the equation for gas flow in the tubing.

39
 Flow Performance in Production Systems M. Stanko

 Compute bottom‐hole pressure from IPR equation.

 Compute separator pressure using bottom‐hole pressure, well rate and pressure loss in tubing and
pipeline.
 Compare if the separator pressure calculated is equal to the given separator pressure, if not, another
well rate is tried.
 The process is repeated until the difference between the given and calculated separator pressure is
minimal.
As an example, consider the production network of two wells shown in Figure 2‐4.

FIGURE 2‐4. PRODUCTION NETWORK WITH TWO WELLS

A methodology for solving the flow equilibrium conditions of the system is the following:

 Assume a surface rate for both wells (q1 and q2)

 Use the inflow performance relationship of each well to calculate the operating bottom‐hole pressures
(pwf1 and pwf2).
 Perform pipe pressure drop calculations from the bottom‐hole of wells 1 and 2 to the wellhead points
p’wh and p’’wh.
 With the separator conditions, the sum of the two liquid rates calculate the wellhead pressure (p’’’wh).
 Iterate on the rate of each well until the three pressures (p’wh, p’’wh, p’’’wh) are the same.
In the following discussions, the flow performance and equilibrium of a production system will be explained
graphically using the available and required pressure curves. These concepts are similar to what is popularly
known as “Nodal Analysis” and are based on the fact that, for a flowing system, the pressure at a given location
must be the same if calculated countercurrent or concurrent from a location with a fixed pressure.
Please note that the graphical method is used just to understand the performance of the production system.
Engineering calculations are made solving the system of equations (i.e. numerical model).

2.1. INFLOW PERFORMANCE RELATIONSHIP

Inflow performance relationship (IPR) expressions are typically derived by solving analytically the partial
differential equations (PDE) of reservoir flow and introducing simplifications and assumptions. The derivation
often yields an expression that relates reservoir and bottom‐hole pressure with reservoir rates at the transient,
steady state and pseudo‐steady state regimes.

40
 Flow Performance in Production Systems M. Stanko

In principle, there should be three independent IPRs, one for each phase that is produced from the formation
(oil, gas and water). However, often the IPR is made for one of the phases (the main phase, oil or gas) and the
other are expressed by using a ratio (gas oil ratio, GOR, water cut, WC). The ratio is often assumed to remain
constant when rate is varied.
Consider the configuration shown in Figure 2‐5. The cross section of a radial reservoir is shown, with a vertical
well drilled in the center. Initially the well is closed so the pressure across the reservoir is constant and equal
to pRo. The well is then open and the wellbore pressure is fixed to pwf.

FIGURE 2‐5. CROSS SECTION OF A VERTICAL WELL DEPICTING THE COORDINATE SYSTEM TO PLOT PRESSURE VERSUS RADIUS

Initially, at time t1 only the vicinity of the well will experience a reduction in pressure because of flow towards
the wellbore (shown in Figure 2‐5). As time passes the pressure will be reduced farther away from the wellbore
until it reaches the boundary re (t3). With time, as the reservoir is depleted and the reservoir pressure falls, the
pressure distribution will continue to change as shown in t4.
The period from t0 to t3 is called infinite acting (or transient) and the period after t3 is called “stabilized flow”
or pseudo steady state (pss), after the pressure changes have reached the outer boundary.

FIGURE 2‐6. EVOLUTION OF PRESSURE ACROSS THE RESERVOIR WITH TIME WHEN PUT ON PRODUCTION

The time required for the reservoir to enter the pseudo steady state depends greatly on the reservoir
characteristics, i.e. permeability, porosity, and properties of the fluid (i.e. viscosity, compressibility). It might
take from a few hours to some years.
In some reservoirs, the pressure is the boundary is kept constant (e.g. due to water injection, aquifer support,
etc.). For those cases, typically referred to as “steady‐state production”, it is assumed that the standard
condition rate is constant for all radial positions.

41
 Flow Performance in Production Systems M. Stanko

An expression to estimate the required time (in hours) to reach pseudo‐steady state or steady‐state is given
in Eq. 2‐1 (for vertical wells and circular drainage area):

𝜙∙𝜇∙𝑐 ∙𝐴
𝑡 281.3 ∙ EQ. 2‐1
𝑘
Where:
𝜙 Porosity [‐]
𝜇 Fluid viscosity at reservoir conditions [cp]
𝑐 7 Total compressibility @ reservoir conditions [1/bar]
𝐴 Drainage area [m2]
𝑘 Permeability [md]

TABLE 2‐1. TIME REQUIRED TO PSS FOR A GAS RESERVOIR WITH THE CHARACTERISTICS8 AND USING EQ. 2‐1.
k tpss k tpss
[md] [h] [md] [h]
0.01 835 552.28 100 83.56
0.1 83 555.23 1 000 8.35
1 8 355.52 10 000 0.84
10 835.55

For reservoirs with medium to high permeabilities, the well enters pseudo steady state or steady state in
relatively short time (minutes, hours, days), thus most of the reservoir outtake will be performed under those
regimes. Moreover, one can frequently remove the time dependence from the equations by relating it to
depletion and using reservoir pressure instead.
For tight formations (k < 1 md) productivity in the transient regime must be considered when estimating the
IPR. Therefore, time typically appears explicitly in these equations, together with an initial pressure.
The IPR typically contains information about:
 Permeability and porosity of the formation
 Well drainage area, formation thickness
 Type of outer boundary – typically no flow or sometimes constant pressure (e.g. if injection from a
neighboring well is being applied)
 Restricted flow to the wellbore (formation damage, stimulation, fracturing, perforation penetration,
gravel pack, screens).
 Wellbore geometry
 Volume‐averaged pressure of the drainage region (reservoir pressure)
 Variation of fluid properties with pressure (viscosity, relative permeability, formation volume factor)
 The convergence effect when fluids flows towards the wellbore.
 Oil, gas and water saturation in the drainage area.

7
𝑐 𝑐 𝑆 ∙𝑐 𝑆 ∙𝑐 𝑆 ∙𝑐
8 ɸ = 0.3 , μ = 0.012 cp, ct = 7.3E‐03 bar‐1, A = 1 130 973 m2

42
 Flow Performance in Production Systems M. Stanko

When deriving IPRs, it is usually possible to separate all parameters that depend on pressure and all
parameters that depend on geometry and integrate them separately (one in space and one in pressure).
Therefore, most IPR equations typically have the following structure:

EQ. 2‐2
𝑞 𝑈∙ 𝐹 𝑝 ∙ 𝑑𝑝

Where the coefficient U is a function of reservoir rock properties, drainage geometry and non‐ideal
phenomena such as skin, partial penetration, etc. The pressure function F(p) usually depends on fluid
properties and on the relative permeability of the phase.
However, partial differential equations of flow in porous media have explicit solutions only for some cases. If
the partial differential equation is too complex, it usually must be solved numerically (like in a reservoir model).
This makes it less attractive for using it in production calculations.
Having an analytical expression derived from the PDE is of great advantage because it allows:

 Quantifying the contribution and relevance of each parameter to well productivity and take corrective
actions, if relevant
 Finding causes for reduced well performance (e.g. in theory, the well should produce “X”, but in
practice, the well is producing “Y”, why?)
 Conducting well planning and completion design
 Predicting well productivity during the planning phase
 Predicting IPR with depletion by updating the equation parameters
However, it is always necessary to adjust the IPR obtained analytically with test data.
Some examples of inflow performance relationship equations are discussed next.

2.1.1. UNDERSATURATED, VERTICAL OIL WELL

The IPR expression for vertical undersaturated oil wells considering skin, radial drainage area and production
in the pseudo‐steady state regime is:

𝑝𝑅
𝑘∙ℎ 1 EQ. 2‐3
𝑞 𝑟 ∙ 𝑑𝑝
18.68 ∙ 𝑙𝑛 0.75 𝑠 𝑝𝑤𝑓 𝜇 ∙𝐵
𝑟
Where:
𝐵 Oil volume factor [m3/Sm3]
ℎ Reservoir thickness [m]
𝑘 Permeability [md]
𝑞 Oil rate at standard conditions, [Sm3/d]
𝑟 Radius of external boundary [m]
𝑟 Radius of well [m]
𝜇 Oil viscosity [cp]
s Skin [‐]

The product 𝐵 ∙ 𝜇 is often linear with pressure, therefore, it’s integration is equal to the product evaluated
at average pressure 𝑝 0.5 ∙ 𝑝 𝑝 . Then the solution of Eq. 2‐4 gives:

43
 Flow Performance in Production Systems M. Stanko

EQ. 2‐4
𝑞 𝐽 𝑝𝑅 𝑝𝑤𝑓

Where J, the productivity index, is:

𝑘∙ℎ 1
𝐽 𝑟 ∙ EQ. 2‐5
18.68 ∙ 𝑙𝑛 0.75 𝑠 𝜇 ∙𝐵 @𝑝𝑎𝑣
𝑟

Effect of depletion on undersaturated oil IPR

The colored lines and points in Figure 2‐7 shows the IPRs calculated using Eq. 2‐5 for an undersaturated oil
well for reservoir pressures equal to 400, 300, 200 bara and several flowing bottom‐hole pressures ranging
from reservoir pressure to atmospheric pressure. While there are some variations of J due to flowing bottom‐
hole pressure and reservoir pressure, the variation is usually small, and J is often considered constant with
depletion.

400
flowing bottom‐hole pressure, pwf [bara]

350

300

250

200

150

100

50

0
0 200 400 600 800 1 000 1 200 1 400
oil rate, qo, [Sm3/d]
FIGURE 2‐7. IPR PREDICTED BY EQ. 2. FOR UNDERSATURATED OIL WELL AND DIFFERENT RESERVOIR PRESSURES

2.1.2. VERTICAL GAS WELL

Assuming a vertical well with cylindrical drainage area, homogeneous formation, pseudo steady state flow,
skin and considering rate dependent skin (turbulent flow), an analytical and general equation for flow of dry
gas is:

𝑝𝑅
7.63 ∙ 𝑘 ∙ ℎ 1 𝑝
𝑞 𝑟 ∙ ∙2∙ 𝑑𝑝 EQ. 2‐6
𝑙𝑛 0.75 𝑠 𝑇 𝑝𝑤𝑓 𝜇𝑔 ⋅ 𝑍
𝑟
Defining:
𝑝
𝑝
𝑚 𝑝 2∙ 𝑑𝑝 EQ. 2‐7
𝑝𝑠𝑐 𝜇𝑔 ⋅ 𝑍

And applying this definition to Eq. 2‐3 gives:

44
 Flow Performance in Production Systems M. Stanko

7.63 ∙ 𝑘 ∙ ℎ 1
𝑞 𝑟 ∙ 𝑚 𝑝 𝑚 𝑝 EQ. 2‐8
𝑙𝑛 0.75 𝑠 𝑇
𝑟
Turbulent flow can be modeled by e.g. including rate‐dependent skin:

7.63 ∙ 𝑘 ∙ ℎ 1
𝑞 𝑟 ∙ 𝑚 𝑝 𝑚 𝑝 EQ. 2‐9
𝑙𝑛 0.75 𝑠 𝐷∙𝑞 𝑇
𝑟
Or, alternatively, to avoid having the presence of standard conditions gas rate in both sides of the equation,
the IPR equation can be re‐written in the form:
𝑛
𝑞 𝐶∙ 𝑚 𝑝 𝑚 𝑝 EQ. 2‐10

With n accounting for the presence of turbulent (n=0.5) or laminar (n=1) flow.
Effect of depletion on dry gas IPR
There is no need to correct for depletion the dry gas IPR presented in Eq. 2‐9.

2.1.3. SATURATED, VERTICAL OIL WELL

Assuming a vertical well with cylindrical drainage area, homogeneous formation, pseudo steady state flow,
skin and considering rate dependent skin (turbulent flow), an analytical and general equation for flow of
undersaturated and saturated oil with simultaneous flow of gas and water is:
𝑘⋅ℎ 𝑘
𝑞 ̄ 𝑟 𝑑𝑝
18.68 ⋅ 𝑙𝑛 0.75 𝑠 𝜇 ⋅𝐵 EQ. 2‐11
𝑟

Where the pressure function is:

𝑘
𝐹 𝑝 EQ. 2‐12
𝜇 ⋅𝐵

And U:
𝑘⋅ℎ
𝑈 𝑟
18.68 ⋅ 𝑙𝑛 0.75 𝑠 EQ. 2‐13
𝑟

If reservoir pressure is equal or below bubble point pressure ( 𝑝 𝑝 ), a typical assumption (which is not
always valid) to solve the pressure function integral is to consider the product ⋅
linear with pressure below
the bubble point. The pressure function can then be expressed as:
𝑘 𝑘 𝑘 𝑘 𝑝
𝐹 𝑝 ∙ EQ. 2‐14
𝜇 ⋅𝐵 𝜇 ⋅𝐵 𝜇 ⋅𝐵 𝜇 ⋅𝐵 𝑝

Or, equivalently:
𝑝
𝐹 𝑝 𝐹 𝑝 0 𝐹 𝑝 𝐹 𝑝 0 ∙ EQ. 2‐15
𝑝

Therefore, the solution of the pressure function integral will have a linear term in addition to the quadratic
term:
1
𝐹 𝑝 𝑑𝑝 𝐹 𝑝 0 ∙ 𝑝 𝑝 𝐹 𝑝 𝐹 𝑝 0 ∙ 𝑝 𝑝 EQ. 2‐16
𝑝 ∙2

45
 Flow Performance in Production Systems M. Stanko

Expanding terms:
1 EQ. 2‐17
𝐹 𝑝 𝑑𝑝 𝐹 𝑝 0 ∙𝑝 𝐹 𝑝 0 ∙𝑝 𝐹 𝑝 𝐹 𝑝 0 ∙ 𝑝 𝑝
𝑝 ∙2
𝑝 𝑝 𝑝
𝐹 𝑝 𝑑𝑝 𝐹 𝑝 0 ∙𝑝 𝐹 𝑝 0 ∙𝑝 𝐹 𝑝 ∙ 𝐹 𝑝 ∙ 𝐹 𝑝 0 ∙
2 𝑝 ∙2 2 EQ. 2‐18
𝑝
𝐹 𝑝 0 ∙
𝑝 ∙2

Grouping terms by pressure:

𝑝 𝐹 𝑝 𝐹 𝑝 0 𝑝
𝐹 𝑝 𝑑𝑝 𝐹 𝑝 0 𝐹 𝑝 ∙ 𝐹 𝑝 0 ∙𝑝 ∙ EQ. 2‐19
2 2 𝑝

Dividing by 𝐹 𝑝 0 𝐹 𝑝 ∙

2
∙ 𝐹 𝑝 𝑑𝑝
𝐹 𝑝 0 𝐹 𝑝 ∙𝑝
EQ. 2‐20
𝐹 𝑝 0 ∙2 𝑝 𝐹 𝑝 𝐹 𝑝 0 𝑝
1 ∙ ∙
𝐹 𝑝 0 𝐹 𝑝 𝑝 𝐹 𝑝 0 𝐹 𝑝 𝑝

Defining a variable “V”

𝐹 𝑝 0 ∙2
𝑉 EQ. 2‐21
𝐹 𝑝 0 𝐹 𝑝

Therefore:
𝐹 𝑝 𝐹 𝑝 0
1 𝑉 EQ. 2‐22
𝐹 𝑝 0 𝐹 𝑝

Substituting back in the integral of the pressure function:

2 𝑝 𝑝
∙ 𝐹 𝑝 𝑑𝑝 1 𝑉∙ 1 𝑉 ∙ EQ. 2‐23
𝐹 𝑝 0 𝐹 𝑝 ∙𝑝 𝑝 𝑝

Substituting Eq. 2‐23 back in the IPR equation:

𝑘⋅ℎ⋅ 𝐹 𝑝 0 𝐹 𝑝 ⋅𝑝 𝑝 𝑝
𝑞 ̄ 𝑟 1 𝑉∙ 1 𝑉 ∙ EQ. 2‐24
18.68 ⋅ 𝑙𝑛 0.75 𝑠 ∙2 𝑝 𝑝
𝑟

Making 𝑞 ̄ , :
𝑘⋅ℎ⋅ 𝐹 𝑝 0 𝐹 𝑝 ⋅𝑝
𝑞 ̄, 𝑟 EQ. 2‐25
18.68 ⋅ 𝑙𝑛 0.75 𝑠 ∙2
𝑟

The following expression is obtained:

𝑝 𝑝
𝑞 ̄ 𝑞 ̄, 1 𝑉∙ 1 𝑉 ∙ EQ. 2‐26
𝑝 𝑝

Vogel found this same equation using data points generated with reservoir simulator, with V = 0.2.
Using Eq. 2‐21, and assuming V = 0.2, 𝐹 𝑝 0 is then:

𝐹 𝑝 EQ. 2‐27
𝐹 𝑝 0
9
Eq. 2‐25 can then be further simplified:

46
 Flow Performance in Production Systems M. Stanko

10 𝑘
𝑘⋅ℎ⋅ ⋅𝐹 𝑝 ⋅𝑝 𝑘⋅ℎ⋅ ⋅𝑝
9 𝜇 ⋅𝐵 @ 𝐽
𝑞 𝑝𝑅 EQ. 2‐28
̄, 𝑟 𝑟 1.8
18.68 ⋅ 𝑙𝑛 0.75 𝑠 ∙ 2 18.68 ⋅ 𝑙𝑛 0.75 𝑠 ∙ 1.8
𝑟 𝑟

Where J is the single‐phase oil productivity index using properties at current reservoir pressure. It can be
pointed out that 𝑞 ̄ , is equal to the absolute open flow9 of the well, calculated assuming an undersaturated
oil well, divided by 1.8.
The backpressure equation, proposed by Fetkovich, is obtained if one applies V = 0 to Eq. 2‐26. This assumes
the pressure function 𝐹 𝑝 is a straight line between zero (“0”) pressure and reservoir pressure.
𝑝
𝑞 ̄ 𝑞 ̄, 1 EQ. 2‐29
𝑝

with 𝑞 ̄ , :

𝑘
𝑘⋅ℎ⋅ ⋅𝑝
𝑘⋅ℎ⋅ 𝐹 𝑝 ⋅𝑝 𝜇 ⋅𝐵 @ 𝐽
𝑞 𝑝𝑅 EQ. 2‐30
̄, 𝑟 𝑟 2
18.68 ⋅ 𝑙𝑛 0.75 𝑠 ∙2 18.68 ⋅ 𝑙𝑛 0.75 𝑠 ∙ 2
𝑟 𝑟

Where J is the single‐phase oil productivity index using properties at current reservoir pressure.
Eq. 2‐29 can be rearranged to look similar to the backpressure equation for low‐pressure dry gas wells:

𝑞 ̄ 𝐶 𝑝 𝑝 EQ. 2‐31
̄,
With 𝐶

To include high velocity effects (turbulent flow), a quadratic term can be included (e.g. Eq. 2‐32 when using
equation Eq. 2‐29), or alternatively, an exponent n, accounting for the presence of turbulent (n=0.5) or laminar
(n=1) flow (e.g. Eq. 2‐33 when using equation Eq. 2‐29).
𝑝
𝑞 ̄ 𝐵∙ 𝑞 ̄ 𝑞 ̄, 1 EQ. 2‐32
𝑝
𝑝
𝑞 ̄ 𝑞 ̄, 1 EQ. 2‐33
𝑝

The IPR equations shown in this section can be used in two ways:

 If no test or field data is available, estimate IPR using the analytical expression. This will require
geometric information, relative and absolute permeabilities, average oil saturation around the
wellbore, fluid properties, etc. However, be aware that assuming 𝐹 𝑝 is linear with pressure is
sometimes not adequate.
 If test data is available, tune the unknown parameters in the equation (e.g. 𝑞 ̄ , and 𝑉 if using Eq.
2‐26) to match the results of the model to the test values. At least two test points are typically
required. It is often reasonable to use V=0 or V=0.2.

9
AOF, oil rate at standard conditions when flowing bottom‐hole pressure is zero bara.

47
 Flow Performance in Production Systems M. Stanko

Effect of depletion on saturated oil IPR

When the reservoir is being depleted, reservoir pressure will decrease below the bubble point pressure, and
fluid properties and saturations around the wellbore will vary, and there will be simultaneous flows or gas, oil
and water towards the wellbore10. Therefore, the IPR will also change.
In Eq. 2‐26, both V and 𝑞 ̄ , depend on the pressure function 𝐹 𝑝 and must be updated with depletion.
However, in some cases, e.g. when using Vogel’s or Fetkovich’s equations, V is often left constant and only
𝑞 ̄, is updated with depletion.
As seem from Eq. 2‐28 and Eq. 2‐30, 𝑞 ̄ , is often a function of reservoir pressure and the mobility evaluated
at reservoir pressure. Therefore 𝑞 ̄ , is often updated by scaling it by the ratio of new to old mobility and by
the ratio of new to old reservoir pressure, as shown in Eq. 2‐34.
𝑘
∙ 𝑝
𝜇 ⋅𝐵
𝑞 ̄, , 𝑞 ̄, , ∙ EQ. 2‐34
𝑘
∙ 𝑝
𝜇 ⋅𝐵

2.1.4. COMPOSITE IPR: BOTH UNDERSATURATED AND SATURATED OIL

If reservoir pressure is above the bubble point pressure, but the flowing bottom‐hole pressure is below the
bubble point pressure (𝑝 𝑝 𝑝 ), then, a suitable equation is a composite IPR, linear above the bubble
point:
If 𝑝 𝑝 then:
𝑝𝑤𝑓
𝑞 ̄ 𝐽 ∙ 𝑝𝑅 𝑝𝑤𝑓 𝐽 ∙ 𝑝𝑅 ∙ 1 EQ. 2‐35
𝑝𝑅

And using a suitable saturated oil IPR below the bubble point pressure, but using as reservoir pressure the
bubble point pressure. To obtain the oil rate one must add the oil rate obtained from the undersaturated oil
equation after substituting 𝑝 𝑝 . For example, using Eq. 2‐30:

for 𝑝 𝑝 :
𝐽∙𝑝 𝑝 𝑝𝑏
𝑞 ̄ ∙ 1 𝐽 ∙ 𝑝𝑅 ∙ 1 EQ. 2‐36
2 𝑝 𝑝𝑅

2.1.5. FLOW OF ASSOCIATED PRODUCTS IN AN OIL WELL: GAS AND WATER

With depletion, due to reservoir pressure reduction and neighboring injection, the saturation of gas and water
around the wellbore will change and the gas and water mobility will increase or decrease, therefore, changing
the producing GOR and WC.
If there is no gas coning from the gas cap nor water cusping from the water layer, then usually an IPR for the
oil phase is used, and the gas and water rates are calculated with the producing GOR and WC. Therefore, it is
typically assumed that GOR and WC remain constant for a given depletion state (or reservoir pressure) even
though there might be variations with 𝑝 for a given reservoir pressure.

The flow of oil and water is sometimes modeled by using a “compound” liquid IPR equation. The liquid IPR is
derived by writing two separate inflow performance relationships for oil and for water. For the case 𝑝 𝑝 :

10
This is evidenced often by a variation in the producing GOR. Please refer to Appendix G for an expression for GOR
variation with time.

48
 Flow Performance in Production Systems M. Stanko

𝑞 ̄ 𝐽 ⋅ 𝑝 𝑝 EQ. 2‐37
𝑞 ̄ 𝐽 ⋅ 𝑝 𝑝 EQ. 2‐38
Adding these two equations gives:
𝑞 𝑞 ̄ 𝑞 ̄ 𝐽 𝐽 ⋅ 𝑝 𝑝 𝐽 ⋅ 𝑝 𝑝 EQ. 2‐39
Effect of depletion on liquid IPR
A new Jl can be found by using the expression:
𝑘 𝑘
𝜇 ⋅𝐵 𝜇 ⋅𝐵
𝐽, 𝐽, ∙ EQ. 2‐40
𝑘 𝑘
𝜇 ⋅𝐵 𝜇 ⋅𝐵

2.1.6. IPR AND WATER OR GAS CONING

Coning from a gas cap or from the aquifer is usually established when the oil rate produced is greater than
critical oil rate or, equivalently, when the flowing bottom‐hole pressure is reduced below the critical bottom‐
hole pressure. The critical oil rate and critical bottom‐hole pressure will depend strongly on the distance
between the well and the water‐oil contact or the gas‐oil contact, among other parameters such as the vertical
permeability.
Immediately after gas or water breakthrough occurs, the oil rate will be severely reduced (for example, Asheim
reports ca 1/10 reduction in the oil rate in a well with water cusping in the Helder field). However, after some
time the oil rate will stabilize, when the transient coning crest stops changing, and the well will then produce
with a constant GOR or WC.
Analytical models indicate that the value of the stabilized GOR or WC depends on the ratio between the oil
rate and the critical oil rate. Interestingly, the relationship is asymptotic, and there will be an oil rate above
which the WC or GOR won’t change significantly. For example, using the analytical steady‐state model of
Asheim for water coning from the aquifer to a horizontal, undersaturated oil well, the water ratio (𝑓 )
has the upper limit:
ℎ
1
ℎ
𝑓 , EQ. 2‐41
𝑘 ∙𝜇
𝜇 ∙𝑘

Where:
ℎ is the combined height of oil and water layers

ℎ is the height of the oil layer

For cases where the critical oil rate is very low (e.g. with regular production rates 10 times higher) it will be
almost impossible to produce the well without causing coning. However, in those situations the producing WC
or GOR will most likely tend to remain constant despite changes in oil rates. Therefore, for these cases it is
often possible to draw an IPR for oil (or total liquid) and calculate the gas (or water) with the stabilized values
of GOR or WC. An example justifying this assumption is the work presented by Astutik[2‐1] that studied an oil
well with simultaneous coning of water and gas.
For cases where oil production rates are comparable to the critical oil rate (0‐8 times larger) or where the
pseudo steady state coning crest is not yet fully established, it is not possible to draw an IPR for one phase and
find gas or water rates with the GOR and WC. More advanced models (or the use of a reservoir simulator) are
usually required.

49
 Flow Performance in Production Systems M. Stanko

2.1.7. IPRS GENERATED WITH RESERVOIR SIMULATOR

IPRs can also be generated from a reservoir model. An example is the Vogel equation, which was derived from
reservoir simulation results.
One common approach so derive IPR is to perform a numerical multi‐rate well test. However, a disadvantage
of this approach is that when the rate is changed, the reservoir will experience a transient. The rate should be
recorded after the transient period has passed. An alternative procedure is presented by Astutik[2‐1] (for
reservoirs with a short transient regime):

 Perform multiple runs of the reservoir model with different well flowing bottom‐hole pressures.
 Extract from the results of each run the oil, gas and water rates for several pre‐specified reservoir
pressures.
 Group and plot all points that have the same reservoir pressure. This will give you the IPR at that
specific reservoir pressure.
Graphically, the proposed procedure is equivalent to make a horizontal sweep in the IPR plot at constant
flowing bottom‐hole pressure, to collect the points at different reservoir pressures.

FIGURE 2‐8. GRAPHIC ILLUSTRATION OF THE PROCESS TO ESTIMATE IPR WITH A RESERVOIR SIMULATOR ACCORDING TO ASTUTIK
(2012)

2.2. AVAILABLE AND REQUIRED PRESSURE FUNCTION

Consider the pipe shown in Figure 2‐9. The pipe segment has an inlet “1” and an outlet “2”. Assume that there
is a single‐phase fluid (e.g. gas, oil or water) flowing through the pipe with a standard‐condition flow rate qsc
(i.e. a constant mass flow rate). In this setup, there are several calculation possibilities:
1. The outlet and inlet pressures are given so the rate flowing through the pipe can be computed.
2. The inlet pressure p1 and the flow rate are given and it allows to compute the outlet pressure p2,
by performing concurrent pressure loss calculations.
3. The outlet pressure p2 and the flow rate are given and it allows to compute the inlet pressure p1
by performing countercurrent pressure loss calculations.

50
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐9. PIPE SEGMENT

If the inlet pressure p1 is left constant, and the standard conditions rate is increased gradually from zero to an
upper limit, the computed outlet pressures p2 (computed with method 2) will display a monotonic concave
curve behavior like the one shown in Figure 2‐10. This curve is the “available pressure” curve.

FIGURE 2‐10. AVAILABLE PRESSURE AT PIPE OUTLET FOR DIFFERENT FLOW RATES AND FIXED INLET PRESSURE

If p2 is left constant and the rate is varied from zero to an upper limit, the computed pressure p1 (using method
3) will follow a convex curve behavior like the one shown in Figure 2‐11. This curve is the “required pressure”
curve.

FIGURE 2‐11. REQUIRED PRESSURE AT PIPE INLET FOR DIFFERENT FLOW RATES AND FIXED OUTLET PRESSURE

Note that the value of the curves at the origin (when there is no flow) is calculated using the hydrostatic fluid
column only, thus it depends on the height difference between inlet and outlet (zero for this particular case,).
The available and required pressure curves concept can be extended to characterize the performance of
complex parts of a production system (that include pipelines, reservoir, pumps, valves, etc.) and when a
multiphase mixture (oil, gas and water) is flowing. For example, consider the well shown in Figure 2‐12 that
includes flow through porous media from reservoir to well bottom‐hole, then pipe‐flow in the casing and pipe‐
flow in tubing. Using the same logic presented earlier, the inlet to the well is the reservoir pressure (considered
invariable for a given depletion state) and the outlet is the wellhead pressure. The available wellhead pressure
curve (often referred to as wellhead performance relationship) will follow the same trend discussed before.

51
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐12. AVAILABLE WELLHEAD PRESSURE VS PRODUCED RATE

There is usually simultaneous flow of gas, oil and water in the well. The available and required pressure curves
are usually built using the flow rate of the preferred hydrocarbon phase (oil or gas). The gas oil ratio (GOR) and
water cut (WC, water surface rate divided by liquid surface rate) usually remain constant when the oil (or gas)
flow rate is varied when building the curve. This means that available and required pressure curves can be
built using the flow rate of any phase of preference, as the others are easily calculated with the GOR and WC.
If the GOR and WC change when varying the flow rate of the preferred hydrocarbon phase (e.g. due to water
coning or gas cusping) an available wellhead pressure curve has to be constructed separately for each phase.
If a wellhead choke is included in the system and the flow through the choke is in the subcritical range, the
curve is modified as shown in Figure 2‐13. It indicates that, if the same rate is desired, a lower pressure p2 has
to be applied in the choked well case than with the no choke case (i.e. there are more pressure losses in the
system).

FIGURE 2‐13. AVAILABLE WELLHEAD PRESSURE WITH CHOKE INCLUDED VS PRODUCED RATE

The required pressure curve can be computed in the same manner but countercurrent departing from a fixed
downstream pressure point (i.e. separator). The curve shown in Figure 2‐14 shows the pressure that has to be
exerted at the bottom‐hole to flow a given rate through the tubing and pipeline. The curve represents the
compound hydraulic performance of the tubing and pipeline (without considering the reservoir).

52
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐14. REQUIRED FLOWING BOTTOM‐HOLE PRESSURE CURVE VS. PRODUCED RATE

The required pressure curve for simultaneous flow of gas, oil and water in a pipe usually displays the shape
shown in Figure 2‐14. The right part of the curve is a friction‐dominated regime (high liquid and gas velocities)
thus an increase in the flow rates give higher pressure drop. The left part of the curve is a gravity dominated
regime (low liquid and gas velocities). For very low velocities, the gas travels faster than the liquid, reducing
the cross‐section flow area occupied by the gas thus yielding a mixture of the density very similar to the density
of the liquid (1 in Figure 2‐15). As the flow rate increases, the liquid begins to travel faster, reducing its flowing
cross section area thus reducing the mixture density (2 in Figure 2‐15).

FIGURE 2‐15. SCHEMATIC REPRESENTATION OF THE MIXTURE DENSITY VARIATION WITH SLIP BETWEEN GAS AND LIQUID
VELOCITIES

In production systems there is simultaneous flow in pipes of two (oil and gas) or three phases (oil, gas and
water). The amounts (mass flow rates) of oil and gas change along the production system due to the decrease
in pressure and temperature. Usually in oil wells the amount of gas increases due to evolving gas out of solution
and in gas producing systems the amount of liquid increases along the tubing due to condensation. However,
the overall composition and total mass flow rate remains constant along the system starting in the reservoir
near the wellbore to the surface, unless there is commingling of different streams or there are transient
phenomena taking place (e.g. liquid accumulation).
An important part of the pressure drop in a production system occurs in the tubing, thus causing significant
gas liberation from the oil and gas expansion, or, similarly, liquid condensation. In consequence, there are
usually multiple flow patterns (phase distribution in the pipe) along the wellbore with different pressure and
temperature gradients (as shown in Figure 2‐16).

53
 Flow Performance in Production Systems M. Stanko

a) Typical flow patterns along the b) Typical flow patterns along the
wellbore in an oil well wellbore in a gas well
FIGURE 2‐16. TYPICAL FLOW PATTERNS ALONG A WELLBORE AS PRESSURE AND TEMPERATURE DECREASE

COMPLETION BITE: TUBULARS

The conductor, casing and tubing are typically made of pipe sections (tubulars) that are threaded together.
The tubulars used by the oil and gas industry can be of two types:
 API11 tubulars: specified and must comply with standards, recommended practices and bulletins
issued by the American Petroleum Institute (API).
 Non‐API tubulars: designed and manufactured outside API specifications.
API tubulars for casing come in three length ranges: 16‐25 ft, 25‐34 ft and 34‐38 ft. API tubulars for tubing
come in two ranges: 20‐24 ft and 28‐32 ft. A pipe section usually refers to as a “joint”
A tag commonly used to refer to tubing and casing tubulars is shown below:

7’’ ‐ 32# ‐ P‐110 ‐ BTC/LTC

F01 F02 F03 F04

Where the fields F01, F02, F03 and F04 have the following information:
 F01: refers to the diameter (nominal or outer) of the pipe in inches. Diameters up to 4½ in are typically
used for tubing (although higher diameters are often used, specially offshore). Diameters above 4½ in
are typically used for casing.
 F02 refers to the weight per length of the pipe (given in pounds per foot or ppf)

11
API: American Petroleum Institute

54
 Flow Performance in Production Systems M. Stanko

 F03 refers to the grade of the steel (yield strength of the material in 1000 psi).
 F04 refers to the thread connection type of the joint.
The “drift” is another important tubular specification that represents the maximum diameter of a cylindrical
mandrel that can be passed without getting stuck inside the pipe. This is different from the pipe inner diameter
(ID) due to ovalization, which is unavoidable in the manufacturing process. Drift must be taken into account
when sending items through the tubular (completion tools, smaller tubulars, etc.)
Tubulars are joined together either by 1. machining them with a male‐threaded end (pin) and female‐threaded
end (box) or 2. by machining them with male‐threaded ends and using couplings. If using couplings, the
coupling is usually threaded in the factory to one end of the joint before shipped to site (a process known as
bucking).

FIGURE 2‐17. TWO JOINTS OF TUBING JOINED BY A COUPLING, OR AN INTEGRATED JOINT

The joints are threaded together (make‐up) when running in‐hole or before running in‐hole (depending on the
height and load capacity of the drilling rig, 2‐3 joints can be threaded before hoisted and run in hole). There
are several methods to “make‐up” joints, but all of them consists on holding the string section that is inside
the well (box), “stab” the suspended section (pin) into the lower section and rotate the suspended section
until certain torque value is achieved.

Most of the methods to calculate pressure drop in multiphase flow are based on first identifying the flow
pattern with some empirical or analytical criteria and use an associated pressure drop model (derived from
mass and momentum conservation equations complemented with empirical correlations). In general, the
information required to compute the pressure gradient (dP/dL) in multiphase flow at a certain PVT condition
is:
 Local volumetric rates to compute superficial velocities of each phase (volume rate of the phase
divided by pipe cross section area).
 Fluid properties: densities, viscosities, fluid‐fluid interfacial tension.
 System properties: Pipe diameter (tubing or casing), roughness, inclination, wettability of the surface,
entry effects (if any).
Due to the change in local volume rates and in flow patterns along the tubing, flowline or pipeline, it is
necessary to perform the pressure drop calculations of the conduit by discretizing into segments.

55
 Flow Performance in Production Systems M. Stanko

The workflow, for the case of a single conduit transporting a standard flow rate of oil, gas and water and where
the temperature of the fluid is known in advance, is the following:
 Discretize the conduit into segments.
 Define a starting point where p0 and T0 is known.
 Calculate local volume rates:
o If using a compositional approach: 1) calculating total mass flow rate, 2) using a PVT model to
calculate fluid properties at P and T.
o If using a Black Oil (BO) approach: 1) converting from standard to local conditions using BO
properties at P and T.; 2) using BO correlations or tables to determine other properties
required (densities, viscosities, etc.).
 Compute superficial velocities
 Estimate pressure gradient (dP/dL c) at the starting point using a multiphase flow model.
 Calculate the pressure in the next point in the conduit by solving numerically the equation Eq. 2‐42 at
the initial conditions p0 and T0.

𝑑𝑝 EQ. 2‐42
𝑐
𝑑𝐿

The numerical method to solve the equation may be explicit or implicit. An explicit 4th‐order Runge‐
Kutta is suggested by the author.
 If the temperature is not given a priori and rather a temperature drop model is available, the numerical
algorithm solves two functions simultaneously, one for pressure and one for temperature.

2.3. FLOW EQUILIBRIUM IN PRODUCTION SYSTEMS

2.3.1. SINGLE WELL PRODUCTION SYSTEM

The production system usually has two boundaries where the pressure is fixed: reservoir pressure and
separator pressure. To find the operating point, the following procedure is followed:

 Select a point of interest in the system

 Compute the available pressure curves considering the system upstream the point of interest down
to the boundary node and
 Compute the required pressure curve considering the system downstream the point of interest up to
the boundary node.
 Intersect the curves to find the operating flow rate.
Figure 2‐18 shows the results of the process for a single well – separator system selecting the wellhead as the
point of interest in the system. The available pressure curves include the pressure losses in reservoir and
wellbore, while the required pressure curve includes the pressure losses in the pipeline keeping separator
pressure constant.

56
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐18. EQUILIBRIUM FLOW RATE OF THE SYSTEM CALCULATED BY INTERSECTING THE AVAILABLE PRESSURE CURVE
CALCULATED FROM RESERVOIR AND THE REQUIRED PRESSURE CURVE FROM SEPARATOR

The production system often contains adjustable equipment such as chokes, ESPs, jet pumps, gas lift, Inflow
control valves (ICV), that can operate at multiple operational settings (e.g. choke opening, ESP frequency, gas
rate, valve opening). The settings of such equipment affect the available or required hydraulic performance of
the system, thus the intersection point of the two curves. Figure 2‐19 shows how the operating rate is reduced
if the choke is fully open or 75% open.

FIGURE 2‐19. EQUILIBRIUM FLOW RATE OF THE SYSTEM FOR: FULLY OPEN CHOKE AND 75% OPEN CHOKE

Hydraulic equilibrium analysis can also be used for design purposes to determine the pressure difference that
an adjustable equipment has to provide to achieve a specific rate. The analysis is carried out by removing the
element from the system and defining the point of interest in the position where the element was. For
example, in Figure 2‐20 an adjustable choke is considered for installation in the system presented. If a rate
below the natural intersection of the curves is desired, the graph allows to estimate the choke pressure drop
required to achieve that rate.

57
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐20. EQUILIBRIUM ANALYSIS EXCLUDING THE WELLHEAD CHOKE TO ESTIMATE CHOKE PRESSURE DROP TO ACHIEVE A
SPECIFIC FLOW RATE

This approach is useful also for ESP and general boosting design (e.g. selecting the inlet and outlet to the pump
as the points of interest, e.g. Figure 2‐21).

FIGURE 2‐21. EQUILIBRIUM ANALYSIS EXCLUDING THE ESP TO ESTIMATE ESP PRESSURE BOOST TO ACHIEVE A SPECIFIC FLOW
RATE

This type of analysis is also relevant for some components that have a numerical model with poor predictability
or with big uncertainties (e.g. multiphase boosters), in which case including it in the numerical model of the
production system might give wrong results.
Please note that this approach does not allow to calculate the adjustable element setting (in the particular
example choke opening) required to achieve the aforementioned pressure difference. For that, the
performance curves of the equipment have to be used.
If a particular equipment is already available (e.g. installed in the well) or selected, then the performance
curves are employed to verify if it is feasible to achieve the delta pressure and rate combination and to
estimate the setting (choke opening or pump frequency) required to achieve that combination. If the operating
condition is not feasible, the operating rate has to be modified.
If there is no particular equipment available or already installed in the well then, a screening is performed
among commercially available equipment to determine which one delivers the required delta pressure and
rate combination. The selection is made taking into account future changes in operating conditions, flexibility
of the equipment, cost, among others.
The required and available pressure curves change with reservoir depletion and in consequence, the pressure
difference required to produce the specified rate changes with time. In Figure 2‐22, the required delta

58
 Flow Performance in Production Systems M. Stanko

pressure across the choke diminishes with time until the desired rate qSC1 is no longer feasible (a negative
choke pressure drop is required, i.e. the choke has to be replaced by a booster).

FIGURE 2‐22. EQUILIBRIUM ANALYSIS EXCLUDING THE CHOKE TO ESTIMATE CHOKE PRESSURE DROP TO ACHIEVE A SPECIFIC FLOW
RATE FOR DIFFERENT TIMES

2.3.2. OPERATIONAL ENVELOPE: CHOKE

A positive (fixed) choke or an adjustable choke at a given fixed opening will display the performance curve
(pressure drop vs. rate) shown in Figure 2‐23. Note that the inlet pressure, the GOR, WC are kept constant.
The rate plotted is the surface rate of the preferred phase (e.g. oil or gas).

FIGURE 2‐23.PERFORMANCE CURVE OF A CHOKE WITH FIXED OPENING

As expected the pressure drop across the choke increases in a non – linear manner when the rate is increased.
However, there is a point where it is not possible to increase the rate further (i.e. the pressure downstream
the choke does not impact the rate flowing through the choke). This is because the fluid velocity at the throat
of the choke has reached the sonic velocity (Figure 2‐24), thus pressure changes downstream the choke do
not affect the upstream conditions. This occurs typically when the pressure ratio is between 0.5‐0.6.

59
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐24. POSITIVE (FIXED) CHOKE IN CRITICAL REGIME (SONIC VELOCITY REACHED AT THE THROAT)

Figure 2‐25 shows the behavior of pressure along the axis of a bean choke. Note that pressure drops suddenly
when the flow encounters the contraction point. In gas‐dominant flows this sudden pressure reduction can
cause cooling (due to the Joule‐Thomson effect), liquid condensation and ice formation (in the presence of
free water).

FIGURE 2‐25. PRESSURE ALONG THE AXIS OF A BEAN CHOKE

Some choke models are derived by applying the Bernoulli equation between a point upstream the choke and
a point on the throat and assuming there are no friction nor localized losses between these two points12. Due
to the convergence of the flow, the effective cross‐section area at the throat is not exactly equal to the throat
cross section, thus a correction factor is introduced that is typically estimated using experimental data. The
pressure measured downstream the choke is usually employed to approximate the pressure at the throat,
assuming there is very little pressure recovery after the throat.
Figure 2‐26 shows the performance curve of the choke when the inlet pressure is varied. The pressure drop at
which the critical flow is reached increases proportionally with the inlet pressure: Δpc ≈ pin – 0.5 pin = pin 0.5.

12
An example of such equations for liquid and gas are derived in appendix B

60
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐26. PERFORMANCE CURVE OF A CHOKE WITH FIXED OPENING

Changes in GOR and WC give a similar variation of the performance curve.

If the choke is adjustable, each choke opening will generate a curve like the one shown in Figure 2‐23. A smaller
opening will provide a larger pressure drop than a larger opening and critical flow will be reached at lower flow
rates. The operational envelope of the choke for multiple choke openings is shown in Figure 2‐27.

FIGURE 2‐27. PERFORMANCE CURVE OF AN ADJUSTABLE CHOKE FOR SEVERAL CHOKE OPENINGS

Some fictitious “desired” operational conditions have been plotted on Figure 2‐27 (with the same inlet
pressure). Points 2 and 3 are feasible, as they fall in the center of the operating envelope of the choke. Point
2 will be operating in the critical range while point 3 will be operating in the subcritical range.
Point 4 falls outside of the choke envelope, which indicates that a larger choke is required for the application.
Point 1 falls in the region of very small choke openings, thus it might be difficult to precisely achieve those
operational conditions. A smaller choke should be considered for this application.
Note that, in Figure 2‐27, for a constant pressure drop, there is an almost linear relationship between choke
opening and flow rate. In reality, this relationship depends on the type of adjustable choke. Figure 2‐28 shows,
for adjustable chokes operating in the subcritical range, the flow rate through the choke (normalized by the
flow rate at maximum opening) versus choke opening. The pressure drop and inlet pressure are kept constant.

61
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐28. FLOW RATE ACROSS DIFFERENT TYPES OF ADJUSTABLE CHOKES WITH A FIXED PRESSURE DROP AND INLET PRESSURE

Chokes that follow the red curve belong to the type “quick opening” (e.g. needle and seat choke). In the low
flow rate range, these chokes are very sensitive to the opening, making it difficult to achieve with accuracy the
desired flow rate. They are however better at higher flow rates.
The black curve corresponds to “equal percentage” chokes (e.g. cage‐type). In contrast with “quick opening”
these chokes have a very good resolution for smaller openings but worsen for higher openings.
The green curve corresponds to linear type chokes. Disk chokes (Willis type) are normally near linear.
Consider the situation shown in Figure 2‐29. A hydraulic equilibrium analysis is performed at the wellhead to
determine required choke delta pressure to deliver a specific rate. The analysis is performed at two depletion
times, 1 and 2. The required pressure curve remains constant with time but the available pressure curve is
reduced due to the decrease in reservoir pressure. The pressure at the inlet to the choke and the choke delta
pressure are estimated from the curve.

FIGURE 2‐29. WELLHEAD EQUILIBRIUM ANALYSIS FOR CHOKE DESIGN FOR TWO DEPLETION STATES

Figure 2‐30 shows the performance curve of a choke that has been selected for the application for the two
inlet pressures given p1 and p2 and for multiple choke openings. Changes in choke opening have a greater
impact in the performance curve than changes in inlet pressure.
Operating points 1 and 2 have been plotted on the figure (Δp and rate). At the operating conditions of point
1, the green performance curves are applicable. The choke is operating in the subcritical range and the choke
opening required to achieve the given delta p is a value less than 60%. At the operating conditions of point 2
the blue performance curves are applicable. The choke is operating in the subcritical range and the choke
opening required to achieve the given delta p is a value close to fully open (100%).

62
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐30. ADJUSTABLE CHOKE PERFORMANCE CURVE FOR DIFFERENT CHOKE OPENINGS AND TWO INLET PRESSURES

2.3.3. BOOSTER
Boosters can be roughly split into two categories: in‐well and out‐of‐well boosters. Within in‐well boosters we
have for example:

 Electric submersible pumps (ESP): multi‐stage centrifugal pumps that are suitable for handling liquids
or liquids with low amounts of gas. These pumps will be explained more in detail in section 2.3.4.
 Jet pumps: these are pumps without moving parts, consisting of a pipe surrounding a nozzle inside
which a power fluid is accelerated, lowering the pressure and sucking the fluid in the annular space.
These pumps are suitable for liquids, mixtures, or gasses. These pumps will be explained more in
detail in section 2.3.6.
 Rod pumps: These are piston pumps where the periodic axial stroking of a piston moving up and down
a chamber will sequentially open a valve located at the bottom to allow flow of fluids into the chamber
and open a valve located at the top of the chamber to push fluids out of the chamber. These pumps
are suitable for liquids.
 Progressive cavity pumps (PCP): These consist of a helical rotor that rotates inside a rubber (or metal)
stator, where the movement of the rotor creates a chamber that moves from inlet to outlet, carrying
the fluid. These pumps are usually suitable for liquids.
Within out of well boosters there are, for example:

 Single phase pumps: these are typically single or multi‐stage centrifugal pumps, that are suitable for
handling liquids or liquids with low amounts of gas. These pumps typically require a separation station
upstream to remove the gas.
 Helico‐axial pumps: these are typically multi‐stage axial or semi‐axial pumps. In these pumps the rotor
blades are long and oriented almost perpendicular to the shaft, which promotes a good mixing of the
phases and inhibits separation. These pumps are suitable for gas‐liquid mixtures up to 80‐90% gas
volume fraction at the inlet.
 Twin‐screw pumps: Typically consist of parallel screws. The movement of the screws creates a
chamber that moves from inlet to outlet, carrying the fluid. These pumps are suitable for gas‐liquid
mixtures up to 90% gas volume fraction at the inlet.
 Wet gas compressors: they are typically multi‐stage axial counter‐rotating compressors. In this
compressor both the rotor and stator resemble a rotor and rotate in opposite directions. Both have
the dual function of accelerating the fluid and pressure recovery. These pumps are suitable for gas‐
liquid mixtures down to 80% gas volume fraction at the inlet.
 Dry gas compressors: they are typically multi‐stage centrifugal compressors. They handle typically gas
with very low content of liquid. These pumps typically require a separation station upstream to
remove the liquid. These compressors will be explained more in detail in section 2.3.5.

63
 Flow Performance in Production Systems M. Stanko

Despite their differences, working principles and the type of fluids they handle, most of these boosters follow
the same thermodynamic principles. The fluid power required by a boosting process is, using the 1st law of
thermodynamics for open systems, equal to the enthalpy difference between inlet and outlet times the mass
flow.
EQ. 2‐43
𝑊 𝑚∙ ℎ ℎ 𝑚 ∙ ∆ℎ

It is common to express the enthalpy difference ∆ℎ as the enthalpy difference required by an ideal boosting
process divided by the adiabatic efficiency (𝜂 ).

∆ℎ EQ. 2‐44
∆ℎ
𝜂
The ideal boosting process, that requires the least amount possible of energy, with no losses, irreversibilities
nor heat exchange with the environment is the one performed at constant entropy (therefore ∆ℎ ∆ℎ ).
The value of the adiabatic efficiency varies depending on the type of booster and the fluid it handles, but it
can be roughly said that it lies between 0.3‐0.8.
This equation indicates that to calculate the real compression power required, one should first compute the
isentropic enthalpy difference and then compute the real enthalpy difference using the adiabatic efficiency.
Consider a situation where the fluid composition is known, the inlet pressure and temperature conditions are
known, and the mass flow is known. Moreover, consider that the required outlet pressure is given (e.g. it has
been estimated with a flow equilibrium analysis). With the composition, the inlet pressure and temperature,
it is possible to compute the inlet specific enthalpy and entropy of the mixture. With this data, the outlet
isentropic enthalpy can be calculated using the same inlet entropy and the required outlet pressure. With the
isentropic enthalpy difference, and the adiabatic efficiency, one can then calculate the required boosting
power. The outlet temperature can be computed by calculating first the actual enthalpy difference (using Eq.
2‐44), then clearing out the outlet enthalpy and then using the outlet enthalpy and the outlet pressure to
compute outlet temperature.
As an example of the procedure described above, consider the pressure‐enthalpy diagram for Methane shown
in Figure 2‐31. This diagram has isotherms and iso‐entropy lines over‐imposed. Consider a compression
process from 50 bara to 100 bara, and inlet temperature 20 °C. The ideal compression process, that requires
the least amount possible of energy, is the one that leads to a situation where the output has the same entropy
as the inlet (blue point). The real compression process, however, is far from ideal and will lead to higher outlet
temperatures than the isentropic (red point). In the case shown in the figure, is has been assumed that the
compressor has an adiabatic efficiency of approximately 68%, i.e. the real enthalpy difference is roughly 1.46
times the isentropic.

64
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐31. PRESSURE‐ENTHALPY DIAGRAM FOR METHANE DEPICTING AN ISENTROPIC COMPRESSION PROCESS AND A REAL
COMPRESSION PROCESS

The same procedure can be applied when the booster is handling liquid, a mixture of gas and liquid, or a gas.
The only difference is that the thermodynamic behavior of the mixture will be different. Figure 2‐32, Figure
2‐33, Figure 2‐34, Figure 2‐35, and Figure 2‐36 depict the pressure enthalpy diagrams of several fluids ranging
from dead oil to methane, gradually increasing the producing GOR of the mixture. The plots have been made
for a range of pressure and temperatures commonly encountered in the petroleum industry.

FIGURE 2‐32. PRESSURE‐ENTHALPY DIAGRAM FOR A DEAD OIL DEPICTING LINES OF ISO‐TEMPERATURE AND ISO‐ENTROPY

For the dead oil case, where for most of the operating region the fluid is liquid, it can be seen that the iso‐
temperature and the iso‐entropy lines are almost vertical are very close to each other. This indicates that the
ideal pumping process will lead to a minimal increase in temperature, and most of the enthalpy difference is

65
 Flow Performance in Production Systems M. Stanko

due to the increase in pressure. It also indicates that one can achieve great pressure differences without
investing much energy, because the lines have a big positive slope.
For a liquid, the enthalpy difference of two points with the same entropy can be approximated by:

∆𝑝 EQ. 2‐45
∆ℎ
𝜌
Substituting this expression in Eq. 2‐43, to compute required boosting power, gives the familiar expression to
estimate pump power consumption:

∆ℎ ∆𝑝 EQ. 2‐46
𝑊 𝑚 ∙ ∆ℎ 𝑚∙ 𝑞∙
𝜂 𝜂
Where q is the local rate at the inlet of the pump.

FIGURE 2‐33. PRESSURE‐ENTHALPY DIAGRAM FOR AN OIL WITH GOR = 150 DEPICTING LINES OF ISO‐TEMPERATURE AND ISO‐
ENTROPY

When the GOR is increased, it can be seen that the iso‐temperature lines will adopt a negative slope and the
slope of the iso‐entropy lines will be diminished, increasing the distance between them. This indicates that the
ideal boosting process will lead to a significant increase in temperature (more significant the higher the GOR),
and a significant part of the ideal enthalpy difference is due to the increase in temperature, besides pressure.
It also indicates that to achieve great pressure differences more energy will be required. This is not a limitation
of the booster, it is a consequence of the thermodynamic behavior of the fluid as it contains a bigger share of
lighter components.

66
Flow Performance in Production Systems M. Stanko

FIGURE 2‐34. PRESSURE‐ENTHALPY DIAGRAM FOR AN OIL WITH GOR=500 DEPICTING LINES OF ISO‐TEMPERATURE AND ISO‐
ENTROPY

FIGURE 2‐35. PRESSURE‐ENTHALPY DIAGRAM FOR AN OIL WITH GOR=10 000 DEPICTING LINES OF ISO‐TEMPERATURE AND ISO‐
ENTROPY

67
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐36. PRESSURE‐ENTHALPY DIAGRAM FOR METHANE DEPICTING LINES OF ISO‐TEMPERATURE AND ISO‐ENTROPY

Exercise: Estimate required boosting power for different fluids

Consider that one needs to increase the pressure of 7 kg/s of fluid from 50 to 100 bara. The inlet temperature
is 50 °C. Consider a fixed adiabatic booster efficiency equal to 0.6, and use the diagrams presented in Figure
2‐32, Figure 2‐33, Figure 2‐34, Figure 2‐35, and Figure 2‐36. Calculate, for each fluid, the outlet temperature
of the fluid, the outlet temperature of the fluid considering an isentropic expansion, and the total required
boosting power.
We follow the steps (for each fluid):
 With the inlet temperature and inlet pressure locate the inlet point on the p‐h diagram and read the
inlet enthalpy.
 With the outlet pressure and assuming the output entropy is equal to the inlet entropy, read the outlet
isentropic enthalpy from the p‐h diagram.
 With the inlet enthalpy and the outlet isentropic enthalpy, calculate the ideal enthalpy difference.
 With the ideal enthalpy difference and using Eq. 2‐44, compute the real enthalpy difference.
 With the real enthalpy difference and the mass flow and using Eq. 2‐43, calculate the power required
by the booster.
 With the real enthalpy difference and the inlet enthalpy, compute the outlet enthalpy.
 With the outlet enthalpy and the outlet pressure and the p‐h diagram, read the outlet temperature.
The results of this process are shown in Table 2‐2. When the content of gas increases, the power and outlet
temperature increase significantly.

TABLE 2‐2. REQUIRED BOOSTING POWER AND OUTLET TEMPERATURES WHEN BOOSTING 7 KG/S OF SEVERAL FLUIDS FROM 50
°C, 50 BARA TO 100 BARA, ASSUMING ADIABATIC EFFICIENCY OF 0.6
FLUID hin Tout,s hs h Power hout Tout
[kJ/kg] [C] [kJ/kg] [kJ/kg] [kW] [kJ/kg] [C]
Dead Oil ‐2086 51 8 13 87.9 ‐2074 53
GOR=150 ‐2369 58 14 23 164.0 ‐2346 62

68
 Flow Performance in Production Systems M. Stanko

GOR=500 ‐2834 70 34 56 393.3 ‐2778 78

GOR=10000 ‐3850 94 82 136 955.4 ‐3714 113
Methane ‐4661 107 118 196 1371.2 ‐4465 135

Boosters typically have a maximum power capacity, dictated by the electric motor (or gas turbine) driving the
shaft. The booster can be operated at that maximum power capacity or at lower values typically by regulating
the rotational speed. Consider a booster operated at constant power 𝑊 ∗ , and the equation to compute the
power of the booster provided earlier:

∆ℎ EQ. 2‐47
𝑊∗ 𝑚∙
𝜂
Eq. 2‐47 specifies that the isentropic enthalpy difference and the mass flow have an inversely proportional
relationship. This means the same fixed number 𝑊 ∗ can be obtained either by a high mass flow rate and a low
isentropic enthalpy difference, or a low mass flow rate and a high isentropic enthalpy difference. Therefore,
an iso‐power line in a booster performance plot depicting iso‐entropic enthalpy difference versus mass flow
should depict a hyperbola (Figure 2‐37).

FIGURE 2‐37. DIAGRAM OF HS VERSUS MASS FLOW FOR A BOOSTER OPERATING AT CONSTANT POWER OF 1.4 MW AND
ASSUMING A CONSTANT ADIABATIC EFFICIENCY OF 1

However, in rotor‐dynamic boosters such as centrifugal and axial pumps, compressors and multiphase
boosters, the adiabatic efficiency is often a function of mass flow (or local flow rate at the inlet) as the blades
and vanes are designed to handle optimally a specific amount of fluid with specific characteristics. These
boosters have a peak of efficiency at a nominal value of rate above and below which the efficiency is reduced
significantly. Because of this, the plot of iso‐entropic enthalpy difference versus mass flow will not depict a
hyperbola. Moreover, boosters typically have a mass flow operating range, and it is often not possible to
operate outside of it, because it leads to low efficiency and failure. The isentropic enthalpy difference is
strongly correlated with the pressure difference, and the mass flow rate with the local volumetric rate at inlet
conditions. Therefore, the performance curve of pressure difference versus local volumetric rate at inlet
conditions at constant power often depicts sort of a straight line with negative slope, due to the concave
behavior of adiabatic efficiency with local rate at inlet.

69
 Flow Performance in Production Systems M. Stanko

2.3.4. OPERATIONAL ENVELOPE: ELECTRIC SUBMERSIBLE PUMP

An electric submersible pump (ESP) operating with undersaturated oil, will display the performance curve
shown in Figure 2‐38. The performance curves have been plotted for several rotational speeds (f). The plot is
made considering that the GOR, WC and fluid viscosity are constant. The maximum rotational speed is f1 and
the minimum rotational speed is f5. For a fixed rotational speed if the rate is increased, the pressure boost
provided by the pump diminishes.
Three lines have been drawn on the plot. To avoid decreased pump life, the pump should always operate
between the minimum (down‐thrust) and the maximum (up‐thrust) lines. When the pump operates outside
this range, there is poor efficiency and excessive wear in the washers (lower or upper) that support the pump
impeller13. The line in violet color is the best efficiency line (BEL) and it is usually desirable to operate as close
as possible to it. The hydraulic efficiency will typically be reduced if the rate is increased or decreased from
this value.

FIGURE 2‐38. PUMP PERFORMANCE CURVE, DELTA PRESSURE VS LOCAL FLOW RATE

The feasible operating region of the ESP is defined by the up‐thrust and down‐thrust lines and the minimum
and maximum rotational speed (which gives a trapezoidal‐like operational region). Besides this operational
constraint, additional constraints that are typically imposed on ESP operation are:
 The suction pressure to be above the bubble point pressure of the crude (often with a safety margin
to avoid vaporization of the crude at the ESP inlet due to inlet losses). If this is not possible, a gas‐liquid
separator should be installed at the inlet and the gas should be produced through the annulus. In
some cases, it is also possible to use impellers that are gas‐tolerant and can be used for inlet gas
volume fractions of up to 10%.
 The total power required (as given by Eq. 2‐48) must be equal or less than the total capacity of the
motor.
Δ𝑝 ∙ 𝑞
𝑃 EQ. 2‐48
𝜂 ∙𝜂
Where:

13
In floating‐impeller configurations, the shaft induces impeller rotation through a key
(https://en.wikipedia.org/wiki/Key_(engineering)) , but the impeller is free to move axially (up or down) with respect to
the shaft.

70
 Flow Performance in Production Systems M. Stanko

𝑃 Required hydraulic pumping power [W]

Δ𝑝 Pressure increase provided by the pump, [Pa]
𝑞 Total volumetric rate at pump inlet conditions [m3/s]
𝜂 Hydraulic efficiency [‐]
𝜂 Mechanical efficiency [‐]

Viscosity greatly affects the performance (and hydraulic efficiency) of an ESP. Figure 2‐39 shows how the ESP
performance map changes when operating with water versus operating with a crude of 200 cP.

FIGURE 2‐39. PUMP PERFORMANCE CURVE OPERATING WITH WATER OR WITH AN OIL OF 200 CP. PREDICTED WITH THE
METHOD DESCRIBED IN THE STANDARD ANSI/HI 9.6.7‐2010

When pumping oil‐water mixtures, the effective viscosity of the mixture will depend on (among other things)
the volume fraction of each phase. Therefore, in wells producing oil and water the viscosity will change with
time when the water cut increases. It is important to take into account this time variation when selecting a
suitable ESP pump for the application.
Density variations will also affect the pressure boost provided by the ESP. Although the operational map in
terms of pump head (ΔH) is unaffected by density, the pressure boost is Δp = ΔH∙ρ , thus it will change with
changes in density.
Consider the situation shown in Figure 2‐40. A hydraulic equilibrium analysis is performed at the pump suction
and discharge to determine required pump delta pressure to deliver a specific rate at two depletion times, 1
and 2. The pressure at the inlet to the pump and the choke delta pressure are estimated from the curves.

71
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐40. ESP EQUILIBRIUM ANALYSIS FOR ESP DESIGN FOR TWO DEPLETION STATES

Figure 2‐41 shows the performance curve of an ESP that has been selected for the application for several
rotational speeds. Operating points 1 and 2 have been plotted on the figure (Δp and rate). The ESP should
operate at a frequency between f4 and f3 to produce point 1, and between f1 and f2 to produce point 2. As time
passes the delta pressure required by the pump will increase until the pump frequency reaches its maximum.
At that moment, the rate should be reduced to move the operating point back into the pump operating
envelope.

FIGURE 2‐41. ESP PERFORMANCE CURVE WITH OPERATING POINTS OVERIMPOSED

2.3.5. OPERATIONAL ENVELOPE: DYNAMIC GAS COMPRESSOR

In systems with gas, compressors are sometimes used to provide additional energy to the fluid to overcome
pressure losses in the surface pipe transportation system. The compressors used for this type of applications
are centrifugal or axial compressors. Axial compressors are used for high rates and medium to low pressure
boosts, and centrifugal compressors are used for low to medium rates and high‐pressure boosts.
The pressure‐enthalpy diagram for Methane is shown in Figure 2‐42. Consider a compression process from 50
bara to 100 bara, and inlet temperature 20 °C. The ideal compression process, that requires the least amount
possible of energy, with no losses or irreversibilities is the one performed at constant entropy (blue line in the
plot).

72
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐42. PRESSURE‐ENTHALPY DIAGRAM FOR METHANE DEPICTING AN ISENTROPIC COMPRESSION PROCESS AND THE
OUTPUT OF A REAL COMPRESSION PROCESS

The real compression process, however, is far from ideal and leads to higher outlet temperatures than the
isentropic (see the red point in the figure labelled with “actual output”). The output of the real process can be
approximated using a polytropic process (p vn = const) where the polytropic exponent n represents the
“efficiency” of the compression process. In the extreme when n= k (ratio of specific heats Cp/Cv), the process
is isentropic.
The outlet temperature of a polytropic process can be estimated with the expression below (with T input in
absolute units, such as K):

EQ. 2‐49
𝑇 𝑇 ∙𝑟

For convenience, the polytropic exponent is expressed as a function of the polytropic efficiency ηp, a number
between 0‐1 (0‐100%):

𝑘 1 𝑛 EQ. 2‐50
𝜂 ∙
𝑘 𝑛 1
The polytropic efficiency typically varies as a function of the operational flow rate. For centrifugal compressors,
it typically lies in the range 0.65‐0.80.
The fluid power required by the polytropic compression process is equal to the enthalpy difference between
inlet and outlet times the mass flow.
EQ. 2‐51
𝑊 𝑚∙ ℎ ℎ 𝑚 ∙ ∆ℎ

The enthalpy difference is estimated using Eq. 2‐52:

𝐻 ∙𝑔 EQ. 2‐52
∆ℎ
𝜂
Where:

73
 Flow Performance in Production Systems M. Stanko

𝑔 Gravitational acceleration [m/s2]

𝐻 Polytropic head, [m], given by:

𝑇 𝑛
𝐻 ∙𝑍 ∙𝑅∙ ∙ 𝑟 1 EQ. 2‐53
𝑔 𝑛 1

Where:
𝑇 Suction temperature [K]
𝑍 Average gas deviation factor between compressor discharge and suction
𝑅 Specific gas constant [J/kg K]
𝑛 Polytropic exponent14
𝑔 Gravitational acceleration [m/s2]
𝑟 Pressure ratio pdis/psuc [‐]

The performance map of compressors is typically expressed in terms of the polytropic head versus the actual
volumetric rate at the compressor suction. Figure 2‐43 shows the performance map of a gas compressor. In a
similar way to the ESP, the polytropic head decreases when the rate increases. Higher rotating speeds increase
the polytropic head.

FIGURE 2‐43. PERFORMANCE MAP OF A GAS COMPRESSOR

The map is contrained to the right by the choke line. In this region, the flow in the compressor reaches the
sonic velocity and cannot be increased further. To the left surge becomes an issue. In surge the flow stops
being steady and there is cyclic flow reversal at the discharge. This is because the polytropic head curve drops
at rates below the surge line.
Consider a compressor with fixed suction pressure and a valve located at the exit. In the operational region to
the right of the surge line if one chokes the valve at the discharge, the rate is reduced and the pressure
delivered by the compressor matches the pressure upstream the valve. However, then the rate is reduced
below the surge rate, the compressor is no longer able to deliver this pressure (due to the concavity of the
curve) and the discharge pressure becomes higher than the pressure the compressor can deliver. This causes

14
The compression of the gas in the compressor is approximated assuming that the process is polytropic (p∙vn = const).
The polytropic exponent can be estimated from measured data.

74
 Flow Performance in Production Systems M. Stanko

flow reversal. With the flow reversal, the pressure at the discharge decreases and falls within the compressor
curve again. The rate increases again and the cycle is repeated.
The performance map of a compressor will vary mainly with the inlet temperature, the molecular weight of
the gas (M) and the heat capacity ratio15 (k = Cp/Cv). If the curves at test conditions are known16, the following
expressions can be used to correct them and estimate the performance under different conditions:

𝑘 𝑀 𝑇 EQ. 2‐54
𝑞 𝑞 ∙ ∙ ∙
𝑘 𝑀 𝑇

𝑘 𝑀 𝑇 EQ. 2‐55
𝐻 , 𝐻 , ∙ ∙ ∙
𝑘 𝑀 𝑇
To avoid having a collection of curves each time these variables change, often the operating point is converted
to the test conditions using Eq. 2‐50 and Eq. 2‐51, and are plotted on the test performance map.
Besides the constraint that the operating points must fall on the compressor performance map, there are
some additional constraints that must be considered:
 The required compression power (P) must be less than the power capacity of the driving motor. The
total required power can be computed with Eq. 2‐56.

∆ℎ ∙ 𝑚 EQ. 2‐56
𝑃
𝜂
Where:
𝑚 Mass flow [kg/s]
𝜂 Mechanical efficiency [‐]

 The temperature at the discharge must be kept below a maximum allowable value. This is usually due
to temperature limits in the compressor seals and safeguard the integrity of the downstream piping.
In systems where chemicals are injected (e.g. hydrate inhibitor), the temperature should also be kept
below the vaporization temperature of the chemical, to ensure its effectiveness.
 The pressure at the suction must be above a minimum allowable value (typically due to rotor
balancing, forces on the shaft and to maintain the effectiveness of seals).
2.3.6. OPERATIONAL ENVELOPE: JET PUMP
A Jet pump, or ejector, is a device without moving parts in which a high‐pressure source fluid is injected to the
main fluid stream through a nozzle. Due to the high speed the high‐pressure fluid achieves in the nozzle, its
pressure is reduced significantly and creates a sucking effect. The suction side of the jet‐pump is then
connected to this location (Figure 2‐44). The high‐pressure fluid is usually single‐phase gas or liquid (oil or
water).

15
A empirical expression to estimate k is k= 1.30 – 0.31∙(g – 0.55)
16
Typically, 1 atm and 15.56 oC at the suction

75
 Flow Performance in Production Systems M. Stanko

HP fluid Discharge

Suction
FIGURE 2‐44. SIMPLIFIED SCHEMATIC OF A JET PUMP

A jet pump displays the performance shown in Figure 2‐45 (taken from Beg and Sarshar[2‐7]). The x axis shows
the mass flow ratio between suction and injection fluid, rm. The y axis shows the pressure ratio between
discharge and suction, rp. There are several curves for different values of rpi which is the pressure ratio between
injection fluid pressure and suction. For a given rpi, if one increases the mass flow ratio rm (injects less fluid)
the possible pressure ratio to achieve between discharge and suction is reduced. If one increases the rpi
keeping the mass flow ratio fixed, a higher rp is achieved.
5
pressure ratio between discharge and suction, rp

4.5
rpi = 2 rpi = 6
4
rpi = 10 rpi = 14
3.5
rpi = 18
3

2.5

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
mass flow ratio between suction and injection, rm

FIGURE 2‐45. PERFORMANCE PLOT OF A JET PUMP (TAKEN FROM BEG AND SARSHAR[2‐7])

An equation to represent this performance is:

𝑟 𝑎∙ 𝑟 EQ. 2‐57

Where “a” and “b” are functions of 𝑟𝑝𝑖 (e.g. polynomials).

A procedure to determine the operational conditions of a production system with a jet pump is the
following:

 Excluding the jet pump, compute the available pressure at the suction (using only the system
upstream the pump) and the required pressure at the discharge (using only the system downstream
the pump) to deliver a desired rate q.
 With these values compute 𝑟 and 𝑟 (with the injection pressure known). Read from the chart the
required mass flow ratio rm, using the appropiate curve of 𝑟 .
 Compute the required mass flow of injection fluid. Subsequently, compute again the required
pressure at the discharge of the pump with the desired rate q and the mass flow of injection fluid.

76
 Flow Performance in Production Systems M. Stanko

 The process is then repeated until the required mass flow ratio rm does not change from iteration to
iteration.

2.4. FLOW EQUILIBRIUM IN PRODUCTION NETWORKS

In a production network the operating conditions in one well affect, to some degree, others, therefore all
possible hydraulic interactions have to be accounted for when computing its hydraulic performance.
The graphical procedure is very rarely used to explain equilibrium calculations of a production network. This is
because, for most cases, pipeline available and required pressure curves depend on the sum of rates of
multiple wells, making it difficult to perform a graphical intersection with the single well pressure curves.
Additionally, the inlet and outlet pressure of pipelines are initially unknown (unless the end of the pipe is the
separator), so the available or required pressure curve have to be redrawn for every inlet or outlet pressure
value assumed.
For these reasons the flow performance analysis of production networks is almost exclusively performed using
computerized routines and software. However, most of the observations given for the single well production
system are applicable to the network case.
Consider as an example the case shown in Figure 2‐46 where there is a production system with three wells, a
pipeline and a separator. The point is interest is defined as the junction where the production of the three
wells is commingled. The available pressure curve is calculated for each well from the reservoir to the junction
and the required pressure curve is calculated for the pipeline from the separator to the junction.

FIGURE 2‐46. PRODUCTION NETWORK WITH 3 WELLS. AVAILABLE JUNCTION PRESSURE CURVE FOR THREE WELLS AND REQUIRED
JUNCTION PRESSURE CURVE FOR THE PIPELINE

To deal with the fact that all available and required pressure curves are drawn with different rates (well 1, 2,
3, 4 and pipeline) an iterative process has to be performed to find the equilibrium rate of each well:
1. Assume a junction pressure
2. Read q1, q2, q3 and q4 from the available and required pressure curves.
3. Verify that q1+ q2 + q3 = q4. If yes, the assumed junction pressure is the real operating pressure. Else,
go back to step 1.

77
 Flow Performance in Production Systems M. Stanko

The mass conservation equation at the junction is checked to verify that the operating junction pressure is
physically consistent.
It is also possible to assume an equilibrium rate for each well and then check that the junction pressure is the
same for all wells and pipeline. However, this is a bit more cumbersome as three variables have to be guessed
instead of one.
It is perhaps more practical to regard the production network as the mathematical function shown in Figure
2‐47. The network model takes as input: properties of the production system, settings of the adjustable
elements and provides as output the well rates.

FIGURE 2‐47. DEPICTION OF THE PRODUCTION NETWORK MODEL AS A MATHEMATICAL FUNCTION

As mentioned before, if there are adjustable elements in the production network, the solution will depend on
the settings of such adjustable elements. Consider, for example, the case of two wells and a pipeline presented
in Figure 2‐48.

FIGURE 2‐48. PRODUCTION NETWORK WITH 2 WELLS

If there are no adjustable elements in the system, the network model function provides a unique combination
of rates q1 and q2 as the solution of the network. However, if each well has an adjustable wellhead choke, the
solution of the network model function becomes dependent on the choke settings. Figure 2‐49 presents a plot
where flow rates of well 1 are plotted in axis “y” and flow rates of well 2 are plotted on axis “x”. When both
chokes are open, the unique combination of rates mentioned earlier is obtained (central red point in Figure

78
 Flow Performance in Production Systems M. Stanko

2‐49). When well 2 is closed, well 1 will be producing alone in the network and the network model solution
yields the point in the y axis. When well 1 is closed, the network solution yields the point on the x axis.

FIGURE 2‐49. WELL FLOW RATE SOLUTIONS FOR NO CHOKE, WELL 1 CLOSED, AND WELL 2 CLOSED

If the network is solved repeatedly for multiple combinations of choke openings, it is possible to create a map
shown in Figure 2‐50 that shows which well rate combinations are achievable by choking and which are not.
Figure 2‐50 has been generated for a production system of two gas wells equipped with wellhead chokes and
discharging to a common pipeline and separator.

FIGURE 2‐50. WELL FLOW RATE DOMAIN SOLUTION FOR THE PRODUCTION SYSTEM WITH 2 WELLS

An interactive calculator is available here https://www.desmos.com/calculator/k6yktvzvfu for the reader to

use and visualize the effect of several system parameters on the operational rates (intersection).
Note that there are two small unfeasible operational regions very close to the x and y axis (the green and violet
lines correspond to well standalone production in the network). For example, when choke 1 is left open and
choke 2 is gradually closed (red line), there will be a point when well 2 cannot physically produce anymore
(and its production becomes zero). This is because the rate of well 2 becomes too small when compared with
the rate of well 1. At that point, if the choke of well 1 is closed slightly (i.e. q1 is reduced), then well 2 will be
able to flow again. This is how the black lines in the drawing (bottom and left bounds of the feasibility region)
were calculated.

79
 Flow Performance in Production Systems M. Stanko

The feasible operational region for a system of 3 wells discharging to a common flowline is shown in Figure
2‐51 a, enclosed by 3 surfaces (one per each well). The operating point is where all surfaces intersect. Figure
2‐51 b shows the effect of choking well 1.

a) Wellhead chokes fully open b) Wellhead choke of well 1 = 50 bar, 2 and 3

fully open
FIGURE 2‐51. WELL FLOW RATE DOMAIN SOLUTION FOR THE PRODUCTION SYSTEM WITH 3 WELLS

Keep in mind this peculiarity of production networks. If there are no adjustable elements, or the adjustable
elements have a fixed setting, there is one unique solution to the production network. However, if there are
adjustable elements in the system, there is usually a variety of operational conditions that can be achieved.

2.4.1. SOLVING NETWORK HYDRAULIC EQUILIBRIUM FIXING WELL RATES

As in the case of the single well production system, it is also possible to solve the network fixing a rate in single
or multiple wells and removing some adjustable elements. In this operational mode, the network model is
used to verify if is physically possible to produce the specified rates and to estimate pressure change that the
adjustable element has to provide.
Consider as an example, the two well network shown in Figure 2‐52. Each well is equipped with a wellhead
choke. Assume that the location downstream the chokes is very close to the junction pj, such as all pressures
(downstream choke 1, downstream choke 2 and junction) can be considered identical. Note the chokes are
ignored when performing the analysis.

80
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐52. WELL FLOW RATE DOMAIN SOLUTION FOR THE PRODUCTION SYSTEM WITH 2 WELLS

By fixing the rate, the bottom‐hole pressure of each well can be calculated with the IPR. With the bottom‐hole
pressure and the rate, it is possible to compute pressure drop concurrent up to wellhead (upstream the
chokes). At the same time, with the well rates and separator pressure, it is possible to compute junction
pressure by performing counter‐current pressure drop calculations.
If pwh1 > pj and pwh2 > pj then it is physically possible to produce the rate and the chokes pressure drop can be
estimated. If, on the other hand, the wellhead pressures are lower than junction pressure, the well rate is not
feasible and it has to be reduced to a lower value. To calculate the well feasible rate, the choke delta pressure
has to be set to 0 (pwh = pj) and the hydraulic equilibrium rate calculated.
This type of analysis can also be done with other adjustable equipment such as ESPs, compressors and
boosters.

2.4.2. DOWNHOLE NETWORKS

Networks can also exist in wells, e.g. when wells are multi‐lateral, multi‐layer or multi‐section. As an example,
consider the horizontal wellbore presented in Figure 2‐53. the horizontal section is open hole, drilled in the oil
layer and swellable packers have been installed along the liner to “split” the wellbore in sections. Each section
is equipped with a sliding sleeve, which can be activated to allow or close communication between the annulus
and the inside of the liner.

81
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐53. HORIZONTAL WELLBORE WITH SEVERAL SECTIONS DELIMITED BY PACKERS

The system shown above can be represented by the equivalent line diagram:

FIGURE 2‐54. EQUIVALENT LINE DIAGRAM REPRESENTING A SECTIONED HORIZONTAL WELLBORE

One can use network calculations, just like in surface gathering systems to estimate how much will each section
produce. In the case that the production is uneven (i.e. significantly different from each section), this might
lead to coning from the aquifer. This occurs typically at the sections closer to the heel, because the flowing
bottomhole pressure is lower. To avoid this problem, sometimes an Inflow control Device (ICD) is placed on
the liner wall of the section to increase artificially the pressure drop.

COMPLETION BITE: SLIDING SLEEVE

A sliding sleeve is a pipe section threaded to the tubing that is used to establish or stop communication at will
between the inside of the tubing and the annulus (annular space between tubing‐casing). Sliding sleeves are
typically used to isolate or connect zones in a well.
The figure below shows the main elements of a sliding sleeve. It consists of two concentric pipe sections, the
outer, which is screwed as part of the production string, and the inner, which can move up or down.

82
 Flow Performance in Production Systems M. Stanko

FIGURE 2‐55. GENERIC SLIDING SLEEVE CONFIGURATION

Locking fingers lock on the grooves to hold in the housing the sleeve position. When the sleeve is pressed
downwards or upwards, the locking fingers will contract when coming out of the groove, and will expand after
reaching the next groove as shown in the figure below.

FIGURE 2‐56. DETAILS OF THE LOCKING MECHANISM OF THE SLEEVE

FIGURE 2‐57. DETAILS OF THE LOCKING FINGERS ON THE SLEEVE THAT RETRACT AND EXPAND WHEN RECIPROCATED AXIALLY
INSIDE THE SLEEVE

Sliding sleeves typically have three positions (although they can have more, depending on the application):
open, closed and equalizing. When the sleeve is in the open position, the slots in the sleeve are aligned with
the holes in the housing.

83
 Flow Performance in Production Systems M. Stanko

SHIFTING PROCEDURE
There are several methods to shift sleeves. The most common is by using wireline (slickline). The general
sequence to shift the sleeve using slickline is shown below. In the first step, the shifting tool is lowered into
the sleeve and jarred down until the collet of the shifting tool is locked into the landing profile of the sleeve.

FIGURE 2‐58. SHIFTING SEQUENCE OF A SLIDING SLEEVE USING SLICKLINE

In the second step, the wireline is further lowered until the retractable keys of the shifting tool are deployed
and sit on the slower end of the sleeve. After this, the wireline is jarred up, to displace the sleeve from the
lower position to the upper position (steps 3‐5).

84
Flow Performance in Production Systems M. Stanko

REFERENCES
[2‐1] Astutik, W. (2012). IPR modeling for coning wells. Thesis for the degree of Master of Science.
Norwegian University of Science and Technology. Trondheim.
[2‐2] Crafton, J. Dyal, V. (1976). An Iterative Solution for the Gas Pipeline Network Problem. SPE‐6032.
Annual Fall Technical Conference and Exhibition of the SPE of AIME. Society of Petroleum
Engineers. New Orleans
[2‐3] Golan, M., Whitson, C. H. (1986). Well Performance. Second Edition. Prentice‐Hall Inc
[2‐4] Litvak, M., Darlow, B. (1995). Surface Network and Well Tubinghead Pressure Constraints in
Compositional Simulation. SPE‐29125. Symposium on Reservoir Simulation. Society of Petroleum
Engineers. San Antonio.
[2‐5] Tian, S., Adewumi, M. A New Algorithm for Analyzing and Designing Two Phase Flow Pipeline
Networks Paper 28177 presented at the 1993 AIChE spring National Meeting, Texas. 1993.
[2‐6] Whitson, C. H. (1983). Reservoir Well Performance and Predicting Deliverability. SPE12518.
[2‐7] Beg, N., Sarshar, Sacha. (2014). Engineers’ Handbook on Surface Jet pumps for Enhanced oil and gas
production. Caltec Limited.

85
 Production Optimization M. Stanko

3. PRODUCTION OPTIMIZATION
In the industry, “Production optimization” is a wide term that englobes detecting opportunities to increase
field oil or gas production, cost reduction and implementing solutions to materialize them. The main principle
is to introduce small cost‐effective changes to improve the production system. Roughly speaking, the potential
increase in production that can be achieved by executing optimization lies someplace between 1‐30%.
Some actions that are typically executed to perform production optimization are:

 Detect locations in the system with abnormally high‐pressure loss and flow restrictions
 Verification of equipment design conditions vs actual operating conditions
 Identification and addressing fluid sources that have disadvantageous characteristics (e.g. high water
cut, high H2S content)
 Identify and correct system malfunctions and non‐intended behavior
 Analyze and improve the logistics and planning of maintenance, replacement and installation of
equipment or in the execution of field activities.
 Review the occurrence of failures and recognize patterns
 Calibration of instrumentation
 Identification of operational constraints (e.g. water handling capacity, power capacity)
 Observe and analyze the response of the system when changes are introduced
 Find control settings of equipment that give a production higher than current (or, preferably, that give
maximum production possible)
 Identify bottlenecks
 Identifying and monitoring key performance indicators (KPIs)
Some tools that are typically used for performing these actions are historic measured and reported data,
instrumentation readings, experiences reported by field operators and in‐house or commercial numerical
simulators.
However, formally speaking, the term “optimization” is wrongly used to characterize most of the activities
described above. In mathematics, optimization refers to find or determine the maximum or minimum value of
a function that is dependent on input parameters while honoring constraints. In practical production
optimization checks are seldom made to verify if the corrections and changes introduced in the system are in
fact the “best” possible alternative. Therefore, a better term to use to englobe most of the activities mentioned
above might be “production effectivization”.
In this chapter only one case of production optimization is discussed: finding control values in the system that
maximize an indicator of economic or system health. Some typical indicators are: oil and gas production rates,
revenue, net present value, ultimate recovery, cumulative production, inverse of lifting costs. The optimization
formulation usually includes multiple operational constraints.

3.1. OPTIMIZING A PRODUCTION SYSTEM

An existing or in‐planning production system has usually adjustable elements, design features that are yet
undecided upon or production and drilling schedules that are modifiable in time. It is usually desirable to find
the particular setting or values of such elements that provides the most attractive operational conditions
within the resources available. The definition of “most attractive operational conditions” depends on the
particular application, the field architecture, the resources available, but it is usually to produce maximum oil,
gas production or revenue (usually called the main objective function) while honoring multiple operational
constraints. Some typical constraints are keeping water production within processing capacity, oil and gas

86
 Production Optimization M. Stanko

within sale specs, electrical power available, injection gas capacity. In most cases, constraints can become
limiting factors that impede to reach an optimum value of the objective function.
The search for optimum operational conditions of the production system is typically performed by using
model‐based mathematical optimization. The model must be accurate enough to predict appropriately the
operating conditions of the production system (taking into account uncertainty in measured data). Otherwise,
any analysis based on the model has limited applicability and usability. Intuitively speaking, the numerical
model is a “surrogate” of the real production system and it can be run with multiple variations of its controls
to find better operating conditions than the current.
Figure 3‐1 shows a simplified scheme designed to close the gap between the output of the model and the
measured variables of the real system. The main task of the data assimilation algorithm is to receive the output
from sensors and change parameters in the model until the difference between the output of the model and
the measured variables is minimized. The parameters in the model are typically properties of the network that
have a high uncertainty (e.g. IPRs) or empirical factors employed by the multiphase pressure drop correlations.
The data assimilation algorithm might also quality control and process (clean, average, validate, aggregate)
the data points that come from the sensors.

Production network
Production network (Numerical model)
(Physical system)
Tuning constants
and properties Predicted
with uncertainty output

Measured
output Data assimilation
Sensors
algorithms

FIGURE 3‐1. DATA ASSIMILATION PROCESS FOR THE NETWORK MODEL (ADAPTED FROM BARROS ET AL, 2015[3‐3])

Using the model has the advantage that it can be queried multiple times to get operational data (in most cases)
quicker than with the real system. Additionally, and depending on the type of model (black box or open), it is
possible to have access to the underlying equations which is a requirement for some optimization algorithms.
Production optimization is often executed in three‐time scales: very short term, short‐term (which also
englobes real time) and long‐term (months‐years).
Short‐term optimization often focuses on finding optimum controls for a given point in time (a particular day
or week), often assuming that the system is at a pseudo steady‐state condition (e.g. for a given depletion
state). There is data available to tune the model. Models typically used are well, gathering system and
processing plant. Typical optimization objectives are oil, condensate or gas production, revenue. Typical
optimization variables are: choke and valve opening, gas lift rates, pump frequency, well routing.
In long‐term optimization the objective is to find optimal control values along time that optimize an indicator
compounded in time (e.g. cumulative production, recovery factor net present value of the project). Typical
control variables are well placement, well rates, well status, number of wells, well routing. Models typically
used are reservoir models often integrated with well, network, processing facilities and economic. Models are
usually highly uncertain.
In some production systems there is often a conflict between short‐term and long‐term optimization, because
the short‐term optimization is oblivious to the effect changes can cause in the future on the system. For
example, in an oil‐rim reservoir with an underlying aquifer and ESP‐lifted wells, a short‐term optimization in

87
 Production Optimization M. Stanko

each time step would probably advice to increase the frequency of ESPs, to produce as much as possible.
However, a long‐term optimization will probably give more conservative ESP frequency values because it will
cause early water breakthrough and reduce ultimate recovery.
In very‐short term optimization the time scale is in seconds, minutes and hours. The objective function is
typically to maximize production, revenue, but it can also be to reduce and mitigate fluctuations (for example
inhibit severe slugging in a well by controlling a choke). Typical objective variables are valve opening, gas lift
rates, pump frequency. It can be deployed using transient or steady state models, but it can also be deployed
directly on the physical system.
For some types of production system, e.g. single choked well, or single ESP‐lifted well, finding the optimum
conditions is a trivial process, just opening the choke to its maximum (Figure 3‐2), or increasing the ESP
frequency to its maximum yields maximum oil production. For this type of cases the optimum point is defined
when operational constraints are met (e.g. power capacity feeding the ESP, sand production in the well,
maximum associated water produced etc.).

FIGURE 3‐2. EQUILIBRIUM FLOW RATE OF THE SYSTEM FOR: FULLY OPEN CHOKE, 75%, 50% AND 25% OPEN CHOKE

For other cases, e.g. gas lifted wells, diluent lifted wells, systems with jet pumps, networks with chokes or ESPs,
etc. there is usually a specific setting that yields optimum operational conditions. Three cases are discussed
next to exemplify these situations. Additionally, a long‐term production optimization case is presented in
section 5.2.3.

3.1.1. CASE 1: GAS‐LIFTED WELLS

Figure 3‐3 shows a well with a gas lift valve installed in it. The plot to the right shows: 1) pressures (at bottom‐
hole, located in the tubing directly in front of the discharge of the gas lift valve) required to flow against
separator pressure for several well oil rates and 2) pressures obtained when the fluid flows from the reservoir
to the same location for several well oil rates. The first curve is affected by the amount of gas injected while
the second doesn’t.
When no gas is injected (i.e. the GOR is the formation GOR), the natural equilibrium oil rate is given by the
intersection between the available (green) and required (magenta) pressure curves (please note that the
bottom‐hole pressure is plotted vs. oil rate). However, when gas is injected through the valve, the GOR of the
tubing and pipeline changes, thus changing the required pressure curve and the intersection point.
When gas is injected in the tubing, the density of the flowing mixture is reduced thus yielding less gravitational
pressure losses. However, the velocity of the mixture increases thus yielding more frictional pressure losses.
When low amounts of gas are injected, the reduction of gravitational pressure losses is higher than the

88
 Production Optimization M. Stanko

increment of frictional pressure losses thus yielding a reduction of pressure in the tubing. However, when the
amount of gas injected is higher, the frictional pressure losses are higher than the reduction of gravitational
losses thus yielding an increase of pressure in the tubing. This change of trend is shown in Figure 3‐4.

FIGURE 3‐3. NATURAL EQUILIBRIUM POINT CALCULATED FOR WELL WITH NO GAS LIFT INJECTION AND WITH GAS INJECTION

FIGURE 3‐4. NATURAL EQUILIBRIUM POINTS CALCULATED FOR DIFFERENT AMOUNTS OF GAS LIFT INJECTED

The oil equilibrium rates for several gas lift rates are plotted in Figure 3‐5. This concave curve is called gas‐lift
performance relationship. The maximum oil production is highlighted in red, where the derivative of reservoir
oil production with respect to gas lift rate is equal to zero.

FIGURE 3‐5. GAS‐LIFT PERFORMANCE RELATIONSHIP

If maximum oil production is desired, the optimum gas injection rate is the x component of the red dot in
Figure 3‐5. However, there might be gas capacity constraints such as there is not enough gas capacity to deliver
this rate. For a single well problem, the optimum gas lift rate can be easily found by plotting the gas lift
performance curve and performing a visual inspection.

89
 Production Optimization M. Stanko

COMPLETION BITE: GAS‐LIFT VALVE

Gas lift valves are deployed in a device called gas lift mandrel (as shown in Figure 3‐6) that is threaded to the
tubing. There are two main types of mandrels, retrievable and conventional.

FIGURE 3‐6. MANDREL TYPES USED TO DEPLOY GAS‐LIFT VALVES

In the case a retrievable mandrel is used, gas‐lift valves can be installed and retrieved at will using wireline or
coiled tubing. The process to lock the gas‐lift valve in the mandrel pocket is shown in Figure 3‐7.

90
 Production Optimization M. Stanko

FIGURE 3‐7. LOCKING PROCESS OF THE GAS LIFT VALVE IN THE MANDREL POCKET

The process to retrieve the gas‐lift valve from the pocket is shown in Figure 3‐8.

FIGURE 3‐8. SEQUENCE TO RETRIEVE A GAS‐LIFT VALVE FROM THE MANDREL POCKET

91
 Production Optimization M. Stanko

Consider two gas lifted wells that are producing to a production separator. Consider further that the two wells
are independent from each other such that the reservoir oil production of each well is only a function of the
well’s gas injection rate qo = f(qg,inj). The behavior is approximated with the polynomial expression shown
below:
EQ. 3‐1
𝑞 𝑓 𝑞 , 𝑎∙𝑞 , 𝑏∙𝑞 , 𝑐∙𝑞 , 𝑑∙𝑞 , 𝑒

The values of the coefficients for the two well are taken from Pavlov et al[3‐5] and are shown in Table 3‐1. The
values of gas injection rate are input in 1E03 Sm3/d and the oil rates are in Sm3/d.

TABLE 3‐1. POLYNOMIAL COEFFICIENTS

Coefficient Well 1 Well 2
a [(Sm3/d)‐3] ‐3.9E‐7 ‐1.3E‐7

b [(Sm3/d)‐2] 2.1E‐4 1.0E‐4

c [(Sm3/d)‐1] ‐4.3E‐2 ‐2.8E‐2
d 3.7 3.1
e [Sm3/d] 12.0 ‐17.0

Figure 3‐9 shows a color map of the total oil production for several combinations of gas‐lift rates injected in
wells 1 and 2 (axis x and y respectively). Contour lines are drawn at constant values of total oil rate. The
maximum is achieved when one injects approximately 100 1E3 Sm3/d in both wells.

FIGURE 3‐9. COLORMAP AND CONTOUR LINES OF TOTAL OIL PRODUCTION AS A FUNCTION OF LIFT‐GAS INJECTED IN WELLS 1 AND
2

If there is a limitation in the total amount of gas available to inject, i.e. 𝑞 , , 𝑞 , , 𝑞 , , this
condition will reduce the feasible area of operation of Figure 3‐9. Figure 3‐10 shows lines (in red) of constant
𝑞 , , . The feasible area of operation will therefore be below the line. If the amount of gas available is

92
 Production Optimization M. Stanko

enough to reach the maximum point, then this is the best operating point. If the amount of gas is not enough,
then, for this particular case, the best is when similar gas lift rates are injected in both wells.

FIGURE 3‐10. COLOR MAP AND CONTOUR LINES OF OIL PRODUCTION AS A FUNCTION OF LIFT‐GAS INJECTED IN WELLS 1 AND 2.
CONTOUR LINES OF TOTAL AVAILABLE GAS‐LIFT RATE.

Mathematical procedure to find the optimal gas lift injection rate

The total oil production function (F) is the sum of the individual well oil production (fi). The total number of
wells is N.

EQ. 3‐2
𝐹 𝑞 , ,𝑞 , ,𝑞 , ,… 𝑓 𝑞 ,

F is a multivariate (N) additively separable scalar function. In order to include the limitation on injection gas
available (∑qg,inj ≤ qg,inj TOT), the Lagrange function is created:

EQ. 3‐3
𝐿 𝑞 , ,𝑞 , ,𝑞 , ,… 𝑓 𝑞 , 𝜆∙ 𝑞 , 𝑞 ,

A necessary condition for this function to be maximum is that the elements of its gradient must be equal to
zero (Eq. 3‐4) and when the additional conditions (Eq. 3‐5, Eq. 3‐6, Eq. 3‐7) are met:

𝜕𝐿 𝑞 , ,𝑞 , ,𝑞 , ,… 𝜕𝑓 𝑞 , 𝜕𝑓 𝑞 , EQ. 3‐4
𝜆 0⇒ 𝜆
𝜕𝑞 , 𝜕𝑞 , 𝜕

EQ. 3‐5
𝜆∙ 𝑞 , 𝑞 , 0

EQ. 3‐6
𝜆 0

93
 Production Optimization M. Stanko

EQ. 3‐7
𝑞 , 𝑞 ,

There are two possible solutions:

Solution 𝜕𝑓 𝑞 , All the wells are operating in their maximum. Valid only if there is
𝜆 0, 0
1: 𝜕𝑞 , enough gas available (∑qg,inj ≤ qg,inj TOT)
Solution 𝜕𝑓 𝑞 , All wells are operating at the same gradient in the gas lift
𝜆 0, 𝜆
2: 𝜕𝑞 , performance curve. Valid only if all the gas available is used (∑qg,inj ‐
qg,inj TOT = 0)

Other cases of gas‐lift optimization with different objective variables and constraints are presented in
Appendix H.

3.1.2. CASE 2: TWO GAS WELLS EQUIPPED WITH WELLHEAD CHOKES

Consider the production system shown in Figure 3‐11 with two dry gas wells with wellhead chokes. The
production of the two wells is commingled and sent to a separator through a pipeline.

FIGURE 3‐11. PRODUCTION SYSTEM WITH TWO DRY GAS WELLS

This production system has been discussed previously. The feasible operational region achievable by adjusting
the wellhead chokes is shown in Figure 3‐12.

94
 Production Optimization M. Stanko

FIGURE 3‐12. WELL FLOW RATE DOMAIN SOLUTION FOR THE PRODUCTION SYSTEM WITH 2 WELLS

It is of interest to evaluate if there is a particular choke opening combination that yields maximum total gas
rate. As a first step, the objective function (total rate qT = q1 + q2) is inspected visually by plotting it in a x,y,z
plot (Figure 3‐13) vs q1 and q2 for the flow rate ranges estimated in Figure 3‐12 (0 < q1 < 6.5.105 Sm3/d and 0
< q2 < 1.0.106 Sm3/d). If the feasible operational region is not taken into account, the maximum will be located
where there is an equal rate distribution between wells 1 and 2.

FIGURE 3‐13. TOTAL GAS PRODUCTION AS A FUNCTION OF WELL 1 AND WELL 2 RATES

However, not all rate combinations plotted in Figure 3‐13 are feasible. In fact there is only a limited operational
region achievable by choking (presented in Figure 3‐12). In Figure 3‐14 the bounds that define the feasible
region have been imposed in Figure 3‐13. The maximum total gas flow rate is obtained when the two chokes
are fully open.

95
 Production Optimization M. Stanko

FIGURE 3‐14. TOTAL GAS PRODUCTION PLOTTED ON THE FEASIBILITY REGION

Note that to obtain a maximum different from the open choke condition, the feasibility area has to be skewed
considerably towards one of the wells. See as an example the modified feasible operational region shown in
Figure 3‐15 (where well 2 has a higher deliverability than well 1) and the resulting total gas flow rate function
in Figure 3‐16. Please note that the maximum occurs now when well 1 is almost completely choked and well
2 is fully open.

FIGURE 3‐15. WELL FLOW RATE DOMAIN SOLUTION FOR THE PRODUCTION SYSTEM WITH 2 WELLS. WELL 2 HAS A HIGHER
DELIVERABILITY THAN WELL 1

96
 Production Optimization M. Stanko

FIGURE 3‐16. TOTAL GAS PRODUCTION PLOTTED ON THE FEASIBILITY REGION

Generally speaking, the shape of the bounds of the feasibility region define where the maximum will be
located. In the simple case presented, two gas wells discharging to a common pipeline, the bounds are fairly
linear, thus tending only to three possible solutions:

 Fully open the choke of the most productive well and close the other
 Open fully both chokes
 Maximum total rate is a plateau that can be achieved by leaving open the most productive well and
choking (with any opening) the least productive well
The shape of the bounds of the feasibility region (linear, parabolic etc) depends on the characteristics of the
production system and have to be calculated and taken into account for each specific case.

3.1.3. CASE 3: TWO ESP‐LIFTED WELLS

Consider the network shown in Figure 3‐17 with two ESP‐lifted wells producing with different water cut (the
full details of the system can be found in Stanko and Golan[3‐6]). Well 1 has a producing water cut of 50% and
well 2 has a producing water cut of 90%. The ESPs can be operated at frequencies between 30‐60 Hz and it is
required to find the optimum combination that yields maximum total oil production.

97
 Production Optimization M. Stanko

FIGURE 3‐17. TWO ESP‐LIFTED WELLS WITH COMMON WELLHEAD MANIFOLD DISCHARGING TO A PIPELINE

In order to find graphically the optimum ESP setting, multiple calculations of the hydraulic equilibrium of the
system were made for different combinations of ESP frequency of wells 1 and 2. The results are shown in the
color map of Figure 3‐18. The color scale represents total oil production and the x and y axis represent the
frequency of wells 1 and 2 respectively. It is possible to see that the frequency of ESP 1 has a higher impact in
the total oil production. The best combination of ESP frequencies is f1 = 60 Hz and f2 = 42 Hz.

FIGURE 3‐18. TOTAL OIL PRODUCTION COLOR MAP FOR THE COMPLETE ESP FREQUENCY RANGE OF WELLS 1 AND 2

Automatic search for optimal operating conditions of the production system

The search for optimum operational conditions of the production system is typically performed by using an
optimization technique and not by brute‐force inspection as presented in the cases earlier. The brute‐force
inspection is useful to understand the problem and the interdependence between objective function, variables
and constraints, but in the general case (multi‐variate) it is impossible to create such plots.
There are mainly two types of optimization: parametric (or static) optimization and dynamic optimization. In
static optimization, the same techniques to maximize or minimize a mathematical function (e.g. f(x,y)) are
applied but on the model of the production system. The model can be a long‐term model (e.g. a reservoir
model) or a short‐term model (network and well).

98
 Production Optimization M. Stanko

In dynamic optimization, control techniques are used to optimize a model, the physical system or a
combination of both. Most control techniques work on a time‐step basis, i.e. sequentially reading variation of
variables, computing new settings of adjustable variables and applying them on the system. For example, in
an oil‐gas separator with a level controller, the control loop reads the liquid level, and outputs the valve
opening to apply on the liquid exit line. This logic can be applied in optimization, for example by driving the
derivative of the objective with respect to the variable to zero in time.
Control techniques usually require a transient model. However, a steady state model can also be used, where
the model is evaluated at each time step.
The optimization technique to employ depends on the optimization problem and the characteristics of the
production system. The problems are commonly continuous (the variation of the adjustable element setting
is usually continuous), constrained and non‐linear (behavior of the objective function). Although there are also
linear, integer problems.
Non‐linear optimization methods are roughly classified in two groups: 1) gradient‐based and 2) derivative free
techniques.
Gradient‐based, as its name indicates, are techniques that use gradient information to estimate a search
direction and calculate the next operational conditions to evaluate. In this type of methods, it is timesaving to
have available analytical expressions for gradients. Otherwise (when working with black‐box models) gradients
can be estimated numerically using finite differences, but it is usually inefficient for large systems because it
requires multiple evaluations of the model on each iteration. An animation of a derivative‐based method
(Newton’s method) is available here:
https://demonstrations.wolfram.com/MinimizingTheRosenbrockFunction/
Derivative‐free techniques perform multiple evaluations on the model and use certain some logic to generate
the next operational points to evaluate. The logic employed depends on depends on the method (examples
are: evolutionary algorithms, pattern search, genetic algorithms, etc.) but it typically consists in using the best
solutions found in one iteration to generate new operational points to test in the next iteration.
An animation of the Nelder‐Mead method is available here:
http://195.134.76.37/applets/AppletSimplex/Appl_Simplex2.html
An animation of a genetic algoritm is available here:
https://demonstrations.wolfram.com/GlobalMinimumOfASurface
And an animation of a pattern‐search algorithm is available here:
https://en.wikipedia.org/wiki/Pattern_search_(optimization)
Linear problems use other family of methods like the Simplex algorithm. An animated example of a linear
problem is given here:
http://optlab‐server.sce.carleton.ca/POAnimations2007/Graph.html
An animated example of the Simplex algorithm applied to the linear problem is:
http://optlab‐server.sce.carleton.ca/POAnimations2007/TwoPhaseGraph.html
An animated example of the Simplex algorithm and the branch and Bound algorithm to solve a linear problem
with integer variables is available here:
http://optlab‐server.sce.carleton.ca/POAnimations2007/MILP.html

99
 Production Optimization M. Stanko

Constraints can be included in the optimization by using Lagrange multipliers, barrier functions. Non‐linear
functions can be linearized by applying piece‐wise linear interpolation (split the function in ranges and use a
linear trend line for each range). However, to avoid using logical operators (if‐like statement to check in which
range a variable x is) that are often incompatible with optimization algorithms, additional variables are
included, such as SOS2 (special ordered set of type 2).
Formulation of the optimization problem
An optimization formulation should contain the following elements:
 Objective
 Decision variables
 Constraints
There are usually two ways to “pose” optimization problems: one is to solve both optimization and modeling
simultaneously and the other is to solve them sequentially. The first alternative is more suitable for models
where there is access to the underlying equations and solving computational routines. In these cases it is
usually possible to use gradient‐based methods that require estimations of the Hessian.
The second alternative is more suitable for black box models, where there is no access to the underlying
equations and computational routines. In these models it is usually favorable to estimate gradients numerically
by perturbing the model several times or to use heuristic optimization algorithm.
To exemplify the difference, consider the following case shown in Figure 3‐19 with two ESP‐lifted wells
producing to a common pipeline and separator.

FIGURE 3‐19. TWO ESP‐LIFTED WELLS WITH COMMON WELLHEAD MANIFOLD DISCHARGING TO A PIPELINE.

One possible mathematical formulation to find the hydraulic equilibrium of the system is to solve the set of
equations:
𝑝 𝐹 𝑞 ,𝑓
EQ. 3‐8
𝑝 𝐹 𝑞 ,𝑓
𝑝 𝐹 𝑞 ,𝑞

Or, equivalently, to minimize: EQ. 3‐9

𝜀 𝑞 ,𝑓 ,𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑞 0

100
 Production Optimization M. Stanko

Where:
𝑞 ,𝑞 Rates of wells 1 and 2, unknown variables @ standard conditions
𝑝 ,𝑝 Junction pressure, unknown variable
𝐹 Pressure drop function for well 1 representing the compound pressure change from
reservoir, tubing, pump, tubing and flowline
𝐹, Pressure drop function for well 2 representing the compound pressure change from
reservoir, tubing, pump, tubing and flowline
𝐹 Pressure drop function for the pipeline, representing the pressure loss in the pipeline.
𝑓 ,𝑓 Rotational speed of ESP pumps 1 and 2 respectively.

For given frequencies of the ESP pumps, it is possible to solve the system of equations (e.g. using a Newton
method) and find the equilibrium rates q1 and q2.
Let’s say now that one wishes to find out ESP frequencies for wells 1 and 2 that maximize the separator rate
𝑞 𝑞 , subject to the constraint that 𝑓 , 𝑓 must be within the operational range of (30‐70 Hz).
If one employs the first method described above to solve this problem, then one must use two sequential
loops:

write
f1 f2

q1 Model
𝜀
optimizer
q2 Equations

write read
model solver

read
FIGURE 3‐20. TWO ESP‐LIFTED WELLS WITH COMMON WELLHEAD MANIFOLD DISCHARGING TO A PIPELINE.

Where, in each iteration of the optimization loop, it is necessary to solve the model once, or several times.
If one employs the second method described above to solve this problem, then a way to formulate the
optimization problem is the following:

Maximize: q1 + q2

By changing: f1, f2, q1, q2

Subjected to the constraints: 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑞 0

30 Hz ≤ f1, f2 ≤ 70 Hz

Note that solving the hydraulic equilibrium of the network has been added as a constraint. This means that
any optimal solution found has to be a feasible operating condition in the numerical model of the network.

101
 Production Optimization M. Stanko

This strategy is used often when optimizing production networks. This optimization problem can be solved
with any suitable method, e.g. a gradient‐based method.
Differences in the formulation
The complexity of the optimization problem can sometimes depend on the decision variables chosen. For
example, and using the case presented earlier of two ESP lifted wells in a network, an optimization formulation
of optimizer and model together which provides a non‐linear formulation is:

Maximize: q1 + q2

By changing: f1, f2, q1, q2

Subjected to the constraints: 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑓 𝐹 𝑞 ,𝑞 0

30 Hz ≤ f1, f2 ≤ 70 Hz

The problem is non‐linear, because functions 𝐹 , 𝐹 , 𝐹 . are non‐linear.

However, by using the model, it is possible to compute the feasible operational rate region of the system, by
running all combinations of pump frequencies f1, f2. Consider the results are similar to the one shown in Figure
3‐21.

(0,b)
f1 = min , f2 = max
q2 (c,d)
f1 , f2 = max

Feasible
operating
region (a,0)
f1 = max , f2 = min

q1
FIGURE 3‐21. FEASIBLE OPERATING REGION OF A SYSTEM WITH TWO ESP‐LIFTED WELLS WITH COMMON WELLHEAD MANIFOLD
DISCHARGING TO A PIPELINE.

Therefore, if the values of a,b,c and d are known a priori and if the boundaries are linear, the original non‐
linear optimization problem can be re‐formulated as a linear optimization problem by:

Maximize: q1 + q2

By changing: q1, q2

Subjected to the constraints: 𝑏 𝑑

𝑞 𝑏 𝑞
𝑐
𝑎 𝑐
𝑞 𝑎 𝑞
𝑑

By performing a priori model evaluations, it is possible to remove the non‐linearity.

102
 Production Optimization M. Stanko

3.2. ISSUES HINDERING THE INDUSTRIAL SCALE ADOPTION OF MODEL‐BASED PRODUCTION

OPTIMIZATION

The industry has been somewhat slow in accepting, adopting and implementing model‐bases optimization
solutions. This can be partly attributed to the following reasons:

3.2.1. FOREIGN FROM THE FIELD’S REALITY

Before embarking on a model‐based production optimization project It is always relevant to ask the question:
is optimization really necessary for this particular case? To execute production optimization entails extensive
use of human, computational resources and time so it is always best to be 100% sure that it is strictly necessary
for the particular case.
Also, is it actually possible to change the decision settings? Is the equipment or actuator functional and
available? Am I allowed to operate the control element? Is the actuator response time compatible with the
optimization workflow?
When formulating the optimization problem, it is crucial to deeply understand (as much as possible) the
underlying physical system. Furthermore, to identify the most important variables, objectives and constraints
and to avoid overcomplicating the problem. This might be sometimes difficult to distill properly during
communications between engineers partly due to lack of understanding of the underlying issues. The lack of
subsequent communication between the optimization engineer and the field operator worsens further the
problem.

3.2.2. MODELS UNCERTAINTY

The uncertainty associated with the numerical models is usually high and there is limited confidence on their
results (this is especially applicable to the case of reservoir models). This raises doubts about the applicability
of the optimum operational controls found and creates resistance and skepticism on the side of operators. It
is always good practice to vary the system parameters (e.g. with a probabilistic sampling method) and quantify
the effect it has on the optimum solution.

3.2.3. NON‐SUSTAINABILITY OF THE PROPOSED SOLUTIONS

Some factors that contribute to the lack of sustainability of the solutions are:

 Lack of expertise on the industry side to understand the basics of the solution provided by consultants
or vendors.
 Ease of use. Not understanding the solution added to difficulties using it often lead to abandoning it.
 The usage of self‐programmed surrogate models that are not easily scalable and maintained. Many
engineers often prefer commercial software where their maintenance, upgrade and troubleshooting
are delegated to a third‐party company.
 Lack of ownership by the industrial partner.
In the area of production optimization, sometimes it is very difficult to develop general solutions and platforms
that are suitable for the majority of field cases encountered. That is why field engineers should always
understand to a great degree the optimization solutions provided by external consultants and vendors.

103
Production Optimization M. Stanko

REFERENCES
[3‐1] Alarcon, G. A., Torres, C. F. & Gomez, L. E. (2002). Global optimization of gas allocation to a group of
wells in artificial lift using nonlinear constrained programming. Journal of Energy Resources
Technology124 (4).
[3‐2] Khan Academy (2018) Lagrange Multipliers, examples. Article retrieved from:
https://www.khanacademy.org/math/multivariable‐calculus/applications‐of‐multivariable‐
derivatives/constrained‐optimization/a/lagrange‐multipliers‐examples.
[3‐3] Barros, E.G.D., Van den Hof, P.M.J., Jansen, J.D. (2015). Value of Information in Closed‐Loop
Reservoir Management. Computational Geosciences 20 (3).
[3‐4] Simplex Optimization method. Article retrieved from:
http://www.chem.uoa.gr/applets/AppletSimplex/Appl_Simplex2.html
[3‐5] Pavlov, A., Haring, M., Fjalestad, K. (2017). Practical extremum‐seeking control for gas lifted oil
production. 2017 IEEE 56th Annual Conference on Decision and Control.
[3‐6] Stanko, M., Golan, M. (2015). Exploring the Potential of Model‐Based Optimization in Oil Production
Gathering Networks with ESP‐Produced High Water Cut wells. 2Presented at the 17th International
Conference on Multiphase Production Technology, Cannes, France, 10‐12 June, BHR‐2015‐J2.

104
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

4. FLUID BEHAVIOR TREATMENT IN OIL AND GAS PRODUCTION SYSTEMS

There are two main methodologies to characterize and quantify fluid behavior: Black oil and compositional. In
the Black Oil (BO) approach three phases are considered: oil (liquid phase), gas (gaseous phase) and water
(liquid phase) and a set of variables are employed to relate the volumetric amounts of the phases at standard
conditions with the volumetric amounts at any pressure and temperature condition. The compositional
approach employs an equation of state (EOS) and the molar composition to estimate fluid properties
(determining numerically the number of phases and calculating their properties).
The BO model can be regarded as a subset of the compositional model where only two components are used:
oil and gas (in most cases water can be treated independently assuming that it does not partition in oil or gas).
The BO approach is still the preferred choice of engineers in the petroleum industry because it is more
“practical” (involves tangible amounts measured in the field) and it is faster than performing time consuming
EOS calculations.
In the past, BO properties were generated from correlations developed for particular fluids and applied to
other cases by using tuning parameters (e.g. typical approach used in commercial software for analysis of
production systems). Nowadays, the generalized procedure is to develop an EOS to characterize oil and gas
fluid behavior and then generate BO properties using this EOS. The BO parameters are pre‐computed and
stored in tables that can later be used as needed by engineers (e.g. in simulators).
The typical workflow to characterize a reservoir fluid is roughly as follows:

 Sampling: a representative sample of the producing fluid is taken. This can be done in three typical
locations in the production system:
o Formation/Well bottom‐hole (oil and gas).
o Test Separator (oil and gas separately). They are later recombined depending on the individual
rates.
o Wellhead
 Determine fluid composition: e.g. using gas chromatography.
 Perform laboratory tests
o CCE Constant composition expansion
o DLE differential liberation experiment
o CVD Constant volume depletion
o MSF Multistage separator experiment
 Develop a PVT model. Development of a physically consistent EOS that represents all the laboratory
tests considering the uncertainties associated to each test and the uncertainties in compositions, and
particularly the properties and amounts of heavy components. Pseudo‐components are typically
employed to represent groups of heavy components (e.g. C7 and above). The properties of such
pseudo‐components are adjusted such that they represent properly the original composition.

4.1. THE BLACK OIL MODEL

To simplify the discussion and the introduction of concepts, an initial assumption will be made that the overall
composition of the fluid stream under study is constant. This is generally not true because in a production
system there is mixing of streams with different compositions, the amounts of oil and gas produced by the
wells change in time due to depletion, gas conning, injection, etc. On a later section, once the basic concepts
are discussed, the consideration is removed.
This chapter does not discuss the presence of water (e.g. gas solubility in water) or the definition and
estimation of its BO properties. Please refer to Chapter 9 of Whitson[4‐1] for details on the topic.

105
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

The BO model is based on the situation when oil and gas at local p and T conditions are brought separately to
standard conditions by passing them through the surface process (P) existing in the field (Figure 4‐1).

FIGURE 4‐1. SCHEMATIC REPRESENTATION OF THE FLASHING OF OIL AND GAS AT LOCAL CONDITIONS TO STANDARD CONDITIONS

Where the subscripts are:

𝑔̅ surface gas component
𝑜̅ surface oil component
𝑔 gas phase @ (p,T)
𝑜 oil phase @ (p,T)

Surface oil (Vō) will be generated from gas phase (Vōg) and from oil phase (Vōo) and surface gas (Vḡ) will be
generated from gas phase (Vḡg) and from oil phase (Vḡo). Four BO parameters, p and T dependent, are then
defined to relate these 4 quantities with the local oil (Vo) and gas (Vg) volume and are summarized in Table
4‐1. These BO parameters are not strict thermodynamic properties, as their values depend on the reference
standard conditions employed (in SPE for instance it is 60°F and 1 atm), the surface process and the
composition.

TABLE 4‐1. BO PARAMETERS

BO Variable Definition
𝑉 𝑝, 𝑇
Oil Volume Factor 𝐵 𝑝, 𝑇
𝑉
𝑉 𝑝, 𝑇
Gas Volume Factor 𝐵 𝑝, 𝑇
𝑉
𝑉
Solution Gas Oil Ratio 𝑅 𝑝, 𝑇
𝑉
𝑉
Solution Oil Gas ratio 𝑟 𝑝, 𝑇
𝑉

106
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

These parameters constitute what is known as the “modified” BO formulation. The traditional BO formulation
does not include the Solution Oil Gas Ratio rs (sometimes called rv). This parameter is important when dealing
with volatile oils and gas condensate fluids.
There are other BO properties that are often recorded at p, T such as oil viscosity (μo), gas viscosity (μg), oil‐
gas interfacial tension (σog). These parameters are often required to perform numerical pressure and
temperature drop calculations in reservoir and production systems.
Note that when the fluid is taken to standard conditions through a specific surface process it will always give
the same GOR and the same oil and gas specific gravities at standard conditions (γo and γg) despite the local p
and T conditions. This is the reason why the GOR, γg and γo are often used to characterize a given fluid.
For certain local p, T conditions, there might be only one phase in equilibrium (oil or gas). In that case, the BO
properties of the non‐existing phase are undefined.
Black oil parameters are usually computed using PVT software with a composition, an EOS (properly tuned to
lab data), and the surface process existing on the field (usually described as a series of separators). The
workflow is presented in Figure 4‐2 and is roughly as follows:

 Take an arbitrary number of moles of the seed composition (zi) to p and T conditions and separate the
oil and gas and store the values of local volumes. The oil will have a composition xi and the gas a
composition yi.
 Take the oil and gas separately through the surface process. At the output gather surface oil and all
surface gas and register the standard conditions volumes.
 Compute BO parameters for the combination of p and T.
 Repeat for several combinations of p and T.

FIGURE 4‐2. SCHEMATIC OF THE PROCESS TO GENERATE BO PROPERTIES

The higher and lower limits for p and T depend on what the BO properties are going to be used for. For
simulations of the reservoir and production system, BO properties have to be generated for the highest and
lowest pressure and temperature expected. However, these values might be unknown a priori as they are
usually a result of the calculations or numerical models that use the BO properties.
It is not failsafe to generate BO properties just between the initial reservoir pressure and temperature and the
first stage separator pressure and temperature. For example, when compression or pumping exists on the
field, pressure and temperature might reach values above reservoir pressure (e.g. at the compressor or pump
discharge) or below separator pressure (e.g. at the compressor or pump suction).

107
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

If the pressure of interest p for which it is desired to generate BO properties is within the range of the pressure
values of the surface process, usually all separators (or equipment) that operate above the pressure p are
neglected in the surface process.
BO parameters are usually stored in a set of tables, where each one contains the variation with pressure for a
fixed temperature. Reservoir or production simulators perform linear (or bilinear) interpolation in these tables
to obtain BO properties at specific p and T conditions.
Figure 4‐3 presents a sketch of the variation of the BO properties versus pressure for a fixed temperature (e.g.
Reservoir) of the hydrocarbon mixture shown in Figure 4‐4. At the given temperature, single phase oil will be
formed for pressures equal or greater than the bubble point pressure.
OIL PHASE

Bo Rs o
GAS PHASE

rs Bg g
FIGURE 4‐3. BEHAVIOR OF BO PARAMETERS VS. PRESSURE FOR A FIXED TEMPERATURE

FIGURE 4‐4. PHASE DIAGRAM OF THE HYDROCARBON MIXTURE USED IN FIGURE 4‐3

108
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

For almost all parameters, there is a change of trend in the curve when the fluid changes from a mixture to
single phase oil. The solution gas‐oil ratio (Rs) remains constant as there is no more free gas in the system to
go into the oil. The value of the oil volume factor (Bo) diminishes with pressure above the bubble point due to
the liquid compressibility. The BO properties of the gas (rs, Bg and μg) become undefined for pressures at and
higher than the bubble point pressure, due to the fact that there is no gas phase at those pressure conditions.
Figure 4‐5 presents a sketch of the variation of the BO properties of versus pressure for a fixed temperature
of the hydrocarbon mixture shown in Figure 4‐6. The temperature is greater than the temperature shown used
in Figure 4‐5 such as gas will be formed at pressures equal or greater than the dew point pressure.
OIL PHASE

Bo Rs o
GAS PHASE

rs Bg g
FIGURE 4‐5. BEHAVIOR OF BO PARAMETERS VS. PRESSURE FOR A FIXED TEMPERATURE

FIGURE 4‐6. PHASE DIAGRAM OF THE HYDROCARBON MIXTURE USED IN FIGURE 4‐5

When the fluid changes from a mixture to single phase gas the solution oil‐gas ratio (rs) remains constant as
there is no more free oil in the system to go into the gas. The BO properties of the oil (Rs and Bo) become

109
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

undefined for pressures at and higher than the dew point pressure, due to the fact that there are no volumes
of oil at those pressure conditions.

4.2. VARIATION OF BO PROPERTIES WITH TEMPERATURE

For engineering analysis of production systems, it is important to capture the variation of BO properties with
both pressure and temperature. The variation of the solution gas‐oil ratio (Rs) with pressure is presented in
Figure 4‐7 for three temperatures. The phase envelope of the fluid is presented in Figure 4‐8.

FIGURE 4‐7. SOLUTION GAS OIL BEHAVIOR WITH PRESSURE FOR THREE TEMPERATURES

FIGURE 4‐8. PHASE DIAGRAM OF THE HYDROCARBON MIXTURE USED IN FIGURE 4‐7

When the temperature increases, the bubble point pressure also increases, however the value of the solution
gas‐oil ratio at the bubble point is the same for all temperatures. This is due to the fact that, once the local
conditions correspond to single phase oil, it does not matter what temperature does it have, it will always
liberate the same amount of gas when brought to standard conditions.
At a given pressure, the solution gas‐oil ratio will be higher for a low temperature compared with a high
temperature.
The variation of the Oil volume factor (Bo) and Gas volume factor (Bg) with pressure are presented in Figure
4‐9 and Figure 4‐10 for three temperatures. Remember that, when the bubble point pressure is reached, there
is no more free gas, therefore, Bg is undefined.

110
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

FIGURE 4‐9. OIL VOLUME FACTOR BEHAVIOR WITH PRESSURE FOR THREE TEMPERATURES

FIGURE 4‐10. GAS VOLUME FACTOR BEHAVIOR WITH PRESSURE FOR THREE TEMPERATURES

The variation of the solution oil‐gas ratio (rs) with pressure is presented in Figure 4‐11 for three temperatures.
The phase envelope of the fluid is presented in Figure 4‐12.
For the particular case shown, the dew point pressure decreases with temperature however the value of the
Solution oil‐gas ratio at the dew point is the same for all temperatures. This is due to the fact that, once the
local conditions correspond to single phase gas, it does not matter what temperature does it have, it will always
liberate the same amount of oil when brought to standard conditions.
At a given pressure, the solution oil‐gas ratio will be higher for a high temperature compared with a low
temperature.

111
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

FIGURE 4‐11. SOLUTION OIL‐GAS RATIO WITH PRESSURE FOR THREE TEMPERATURES

FIGURE 4‐12. PHASE DIAGRAM OF THE HYDROCARBON MIXTURE USED IN FIGURE 4‐12

4.3. VARIATION OF BO PROPERTIES WITH COMPOSITION

When a fluid of certain composition is taken to standard conditions through a specific surface process it
produces a unique value of gas‐oil ratio (GOR) and unique values of oil and gas specific gravities at standard
conditions (γg and γo). Therefore, a change in GOR is always a safe indicator of changes in the composition.
Note that the GOR of the stream is always equal to the saturation Rs of the mixture, or to the saturation 1/rs
(depending if the temperature of study is below the critical temperature of the mixture or above, as the cases
shown in Figure 4‐4 and Figure 4‐6).
In the development done in the previous sections the molar composition of the fluid was kept constant when
generating BO properties (this composition typically comes from well samples). However, this is seldom the
case as in a production system there is mixing of streams with different compositions, the well producing GOR
(or the GOR in each cell of the reservoir model) typically changes in time due to depletion, gas conning, gas
injection, etc.
Changes in GOR will affect black oil properties. An example of this is shown in Figure 4‐1317.

17
Here the new compositions for each GOR have been generated using compositions of separator gas and oil and the
procedure highlighted in section 4.6.

112
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

3.5 800.0
GOR = 143
GOR = 535 700.0
3.0
GOR = 1070
600.0
2.5

500.0

Rs [Sm3/Sm3]
Bo [m3/Sm3]

2.0
GOR = 143
OIL PHASE

400.0
1.5 GOR = 535
300.0
GOR = 1070
1.0
200.0

0.5
100.0

0.0 0.0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
pressure, p, [bara] pressure, p, [bara]

Bo Rs
1.0E‐03 2.5E‐02

9.0E‐04 GOR = 143

8.0E‐04 2.0E‐02 GOR = 535
GOR = 1070
7.0E‐04
GOR = 143
rs, [Sm3/Sm3]

Bg, [m3/Sm3]
6.0E‐04 1.5E‐02
GOR = 535
GAS PHASE

5.0E‐04 GOR = 1070

4.0E‐04 1.0E‐02

3.0E‐04

2.0E‐04 5.0E‐03

1.0E‐04

0.0E+00 0.0E+00
0 100 200 300 400 500 600 0 100 200 300 400 500 600
pressure, p, [bara] pressure, p, [bara]

rs Bg
FIGURE 4‐13. BLACK OIL PROPERTIES ESTIMATED FOR DIFFERENT COMPOSITIONS (GOR). NOTE THAT FOR GOR=1070, THE
FLUID IS NOT ANYMORE AN OIL BUT A GAS IN UNDERSATURATED CONDITIONS.

In situations where a variation of GOR is expected, one must generate black oil tables as a function of GOR, in
addition to pressure and temperature. A tri‐linear interpolation is used to find black oil properties for a given
GOR, p and T.
Saturation pressure (bubble point pressure and dew point pressure) are sometimes used instead of GOR.
When a GOR and p and T conditions are provided to estimate BO properties, the first analysis to perform is to
determine if the fluid is in saturated or undersaturated conditions. A robust approach to do this is to
precompute, at constant temperature, for a large range of GORs (i.e. compositions), the bubble point pressure
(or dew point pressure) and the bubble point Rsb (or the dew point rsd, depending if at the given temperature
the undersaturated fluid is gas or oil). The Rsb and 1/rsd are plotted vs pb and pd (Figure 4‐14).

113
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

FIGURE 4‐14. Rs AND 1/rs VS p COMPUTED FOR SEVERAL GORS AT CONSTANT TEMPERATURE

With the given GOR one can enter on the y axis (arrow in blue in Figure 4‐14) and identify if the undersaturated
fluid is gas or oil and read on the x axis the pb or pd. If the given pressure is less than the saturation pressure,
then the fluid is in saturated conditions and saturated properties have to be used. If the pressure is greater
than the saturation pressure, then the fluid is in undersaturated conditions, and undersaturated properties
should be used with the given GOR.
In general, for the following cases:

 The reservoir is undergoing gas injection

 The reservoir is undergoing gas recycling
 The production system commingles production of pay‐zones with different compositions
 The reservoir has compositional heterogeneities (e.g. with depth or region)
It is usually necessary either to use a compositional model or to separate the reservoir or production system
in sections with constant composition and build a BO table for each composition. If an analysis requires mixing
or commingling fluids from different compositions then an equivalent black oil table has to be developed
depending on the mixing ratio or a compositional approach should be used.
However, when the fluid stream is a combination of gas and oil from reservoir oil, or gas and oil from reservoir
gas, it might still be possible to use a common black oil table for all resulting compositions (albeit with some
modifications).
Consider as an example an oil well as the one shown in Figure 4‐15. The well will produce more and more gas
with depletion as the pressure diminishes around the well bore and the gas mobility increases (an expression
to compute the producing GOR is provided in Appendix G). Figure 4‐16 shows how the phase envelope of the
well stream changes as the GOR increases.

114
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

FIGURE 4‐15. OIL WELL

FIGURE 4‐16. VARIATION OF THE PHASE ENVELOPE WITH CHANGES IN COMPOSITION (GOR)

In this case it could be possible to use a unique black oil table for the complete life of the well.
The original composition has an initial Rsi (or rsi, depending on the type of fluid) that is equal to the GOR. The
GOR can change in two possible ways:
 The new GOR (GOR2) is lower than the initial GOR (GOR1). In this case, one can expand the original
fluid composition to the saturation pressure that gives Rs = GORnew (black arrow shown in Figure 4‐17)
and use this composition for the generation of new black oil properties:
o For pressures below this saturation pressure (region marked with the letter “S” in Figure 4‐17),
the saturated BO properties of this composition will be identical to those found with the
original fluid composition, therefore, it is not necessary to estimate them.
o For pressures above this saturation pressure (region marked with the letter “U” in Figure
4‐17), the undersaturated properties will be different from those of the original composition
and must be computed.

FIGURE 4‐17. Rs VARIATION WITH COMPOSITION WHEN THE NEW GOR IS LOWER THAN THE ORIGINAL GOR

115
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

 The new GOR (GOR2) is higher than the initial GOR (GOR1). In this case one must estimate a new
composition for this new GOR to estimate black oil properties. One approach that can be used to do
this is to:
o Extract the composition of the incipient gas at the saturation pressure of the original
composition
o Add this incipient gas to the original composition until the saturation pressure of the resulting
mixture is equal to the critical pressure of the resulting mixture
The saturated black oil properties for GOR values higher than the original mixture GOR are found by expanding
the new mixture between its saturation pressure (critical pressure) and the saturation pressure of the original
mixture. The saturated black oil properties for GOR values below the GOR of the original mixture will be equal
to those of the original mixture and it is not necessary to estimate them.
Figure 4‐18 shows the Rs behavior versus pressure plotted for the composition of the well at an early time “1”
and a later time “2” with higher GOR. The saturated Rs values (for p < pb1) are the same for both times, but the
saturation Rs, (GOR) has changed.


FIGURE 4‐18. Rs VARIATION WITH COMPOSITION WHEN THE NEW GOR IS HIGHER THAN THE ORIGINAL GOR

The undersaturated properties for GOR values between the GOR value of the original and the new mixture are
found by expanding the new mixture to a saturation pressure that gives the desired GOR value, and then using
the resulting composition of the saturation oil and gas to compute undersaturated black oil properties.
A note when using BO correlations
If BO properties are generated using BO correlations, the behavior of properties estimated at other GORs will
be as discussed above.
A comment on BO tables
BO tables (e.g. those used in reservoir simulators) usually have two parts. One provides values of BO properties
in the saturation region (continuous lines in Figure 4‐19) and one that provides BO property values for the
undersaturated region at multiple GORs.

116
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

Rs Bo rs
FIGURE 4‐19. VARIATION OF MAIN BO PARAMETERS WITH COMPOSITION WHEN MORE GAS FLOWS INTO THE WELLBORE

4.4. BO CORRELATIONS
There are many correlations for BO parameters of the oil phase developed for particular fields, fluids and
regions. Typically, these correlations are accurate only if the fluid of interest and pressure and temperature
conditions are similar to those for which the correlations were developed. As an example, Table 4‐1 shows
some expressions for pb, RS, Bo, Bg and rs.

TABLE 4‐2. SELECTED CORRELATIONS FOR BO PARAMETERS

PROPERTY CORRELATION AUTHOR
.
𝑅 . ∙ . ∙
Bubble pressure 𝑝 1.995 ∙ ∙ 10 1.7566 Standing (1977)[4‐2]
𝛾
𝑅 0.571 ∙ 𝛾 ∙ 10 . ∙ . ∙
Gas‐in‐oil ratio . Standing (1977)[4‐2]
∙ 0.797 ∙ 𝑝 1.4
𝐵 0.9759 0.000952
Oil formation .
𝛾 . Standing (1977)[4‐2]
factor ∙ ∙𝑅 0.401 ∙ 𝑇 103
𝛾
Gas formation 𝑇∙𝑍
𝐵 0.00351 ∙ Definition
factor 𝑝
Oil‐in‐gas ratio 𝑟 1.25 ∙ 10 ∙ 𝑅 ∗ ∙ 0.08 4 ∙ 10 ∙𝑝 . Whitson (1994)[4‐3]

Where

𝑝 Fluid pressure [bara]

𝑅 Gas solubility [Sm3/Sm3]
𝑅∗ Gas solubility @ 345 bara [Sm3/Sm3]
𝑇 Fluid temperature [K]
𝑍 Generalized compressibility factor [‐]
𝛾 API gravity [‐]
𝛾 Stock tank oil gravity [‐]

As in the case of the BO tables, the given expressions for RS, Bo, Bg and rs describe the behavior of the
undersaturated region. The first step is then, with the given GOR, calculate the saturation pressure (e.g. bubble
pressure) at the given temperature. If the pressure of interest is below the saturation pressure, the
correlations are used as shown. Otherwise, an expression for undersaturated BO properties has to be used.

117
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

As mentioned earlier, tuning parameters can be introduced in the BO correlations to match test or field data.
This is usually done (in commercial software) by introducing two constants:

𝑣𝑎𝑙𝑢𝑒 ∙ 𝐴 𝐵 𝑣𝑎𝑙𝑢𝑒 EQ. 4‐1

Where:
𝐴 Tuning multiplier
𝐵 Tuning shifting
𝑣𝑎𝑙𝑢𝑒 Calculated property (using correlation)
𝑣𝑎𝑙𝑢𝑒 Measured property

These parameters A and B are changed by an optimization engine to minimize the difference between the test
or field data and the correlation output.

4.5. BO PROPERTIES IN PRODUCTION CALCULATIONS

The most typical use of BO properties in production system calculations is to convert from standard conditions
rates to local rates and vice‐versa. The local rates are used for pressure drop calculations, flow assurance
analysis, among others. Figure 4‐20 presents the relationship between standard conditions rates and local
conditions rates and vice versa using a BO transformation matrix. These expressions assume that water is not
soluble in oil nor gas.

1 𝑅
⎡ 0⎤ 𝐵 𝑅 ∙𝐵
𝑞 ⎢𝐵 𝐵 ⎥ 𝑞 𝑞
⎡
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟
0⎤
𝑞
⎢𝑟 1 ⎥ ⎢ ⎥
𝑞 0⎥ ∙ 𝑞 𝑞 ⎢ 𝐵 ∙𝑟 𝐵 ⎥ ∙ 𝑞
⎢𝐵 𝐵 ⎢ 0⎥
𝑞 ⎢ ⎥ 𝑞 𝑞 𝑞
1⎥ ⎢1 𝑅 ∙ 𝑟 1 𝑅 ∙𝑟 ⎥
⎢0 0 ⎣ 0 0 𝐵 ⎦ ,
⎣ 𝐵 ⎦ ,

Standard conditions calculated from Local conditions calculated from standard conditions
local conditions
FIGURE 4‐20. TRANSFORMATION MATRIXES TO TAKE STANDARD CONDITIONS RATES TO LOCAL CONDITIONS AND VICE VERSA

A similar relationship can be developed between the local and surface condition densities (Figure 4‐21).

𝐵 𝑟 ∙𝐵 1 𝑟
⎡ 0⎤ ⎡ 0⎤
𝜌 ⎢1 𝑅 ∙ 𝑟 1 𝑅 ∙𝑟 ⎥ 𝜌 ⎢𝐵 𝐵 ⎥
𝜌 𝜌 𝜌
⎢ 𝑅 ∙𝐵 𝐵 ⎥ ∙ 𝜌 ⎢𝑅 1 ⎥
𝜌 ⎢ 0⎥ 𝜌 𝜌 0⎥ ∙ 𝜌
⎢𝐵
⎢1 𝑅 ∙ 𝑟 1 𝑅 ∙𝑟 ⎥ 𝜌 ⎢
𝐵
⎥ 𝜌
⎣ 0 0 𝐵 ⎦ , ⎢0 1⎥
0
⎣ 𝐵 ⎦ ,
Standard conditions calculated from local Local conditions calculated from standard
conditions conditions
FIGURE 4‐21. TRANSFORMATION MATRIXES TO TAKE STANDARD CONDITIONS DENSITIES TO LOCAL CONDITIONS AND VICE VERSA

This expression assumes that:

 The density of the surface oil coming from local oil (ρōo) is the same as the surface oil coming from
local gas (ρōg) and

118
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

 The density of surface gas coming from local gas (ρḡg) is the same as surface gas coming from local oil
(ρḡo).

4.6. ESTIMATION OF A NEW COMPOSITION WHEN THE WELL GOR CHANGES

In some occasions, it is desirable to estimate the well or stream composition given a producing GOR, for
example when the producing GOR of the well changes in time and no sampling has been performed or when
reservoir simulations are not available. The new composition can usually determine by recombining some
“source” oil and gas at different proportions (as shown in Figure 4‐22, using the β multiplier). This setup can
be modeled in any PVT calculator and the parameter beta varied until a match is obtained with the new GOR.

FIGURE 4‐22. RECOMBINATION OF SOURCE GAS AND OIL TO YIELD STREAM COMPOSITION

There are two methods generally accepted to obtain source compositions of oil and gas: 1) separator oil and
gas generated after passing a measured composition through the surface process use and 2) Using the last
known measured composition, take it from the reservoir depletion state when it was measured to the current
reservoir depletion state. This should be performed using the process with approximates the best the actual
process that undergoes in the reservoir (e.g. CVD, CCE).
Note that in some occasions the GOR is measured by means of a test separator thus the surface process to
employ must be different than the surface process used during normal production.

4.7. WATER‐RELATED PROPERTIES

In some cases, it is necessary to account for the content of water in gas and gas in water. (e.g. condensation
of water in natural gas production systems or liberation of gas in water‐dominated flow). Two factors are
typically employed for this purpose: rsw (p,T) that provides the water content in gas when the gas is fully
saturated with water at a given pressure and temperature, and Rsw, that provides the gas content in water
when the water is fully saturated with gas at a given pressure and temperature.
Despite often being expressed as a ratio of standard conditions volumes, rsw and Rsw are not calculated
considering the surface process or by taking the saturated mixture of gas and water to standard conditions.
They are calculated by first computing the mole fraction of water in the gas and water mixture (or gas in the
water and gas mixture) and then converting the mole fraction to a volume fraction, with conversion factors
(i.e. molar densities) calculated at standard conditions of pressure and temperature.
Despite this caveat, rsw and Rsw can still be used to compute local rates of water or gas along the production
system, just like black oil properties. For example, consider a gas well, where the gas entering the wellbore is
saturated with water at TR, pR conditions, and there is no free water entering the well. To calculate local rates
of water at any p,T combination in the tubing, one can use the following expression:

119
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

EQ. 4‐2
𝑞 𝑚𝑎𝑥 0, 𝑟 @ , 𝑟 @ , ∙ 𝑞 ∙ 𝐵 𝑝, 𝑇

The difference in rsw could be positive or negative. If positive, it means that the gas has lowered its water
content (capacity of carrying dissolved water) and therefore some water will condense out of the gas as a
liquid phase. If negative, it means that the gas has increased its capacity of carrying dissolved water and could
take more water than what currently has, i.e. there is no water out of solution as a liquid phase. However, as
there is no more water available to go into the gas, a lower limit of zero is imposed to the expression.
When there is free water entering the flow system besides water‐saturated gas, then the equation must be
modified as follows:

𝑞 , EQ. 4‐3
𝑞 𝑚𝑎𝑥 0, 𝑟 @ , 𝑟 @ , ∙𝑞 ∙ 𝐵 𝑝, 𝑇
𝐵 𝑝 ,𝑇

Where inj is the water injection point, and 1 is the conditions at which the gas was saturated with water. The
“max” function is used here again in case all the injected free water vaporizes into the gas, because of the
increase in rsw.
Figure 4‐23 shows a color map of rsw as a function of pressure and temperature. rsw is inversely proportional
to pressure and directly proportional to temperature.

FIGURE 4‐23. COLOR MAP OF AS A FUNCTION PF PRESSURE AND TEMPERATURE

Corrosion or liquid loading but no free water, how is this possible?

In pipe and wellbore flow there could be some situations where there is no free water at the inlet and outlet
of the conduit, but there is free water some place in between. This is because values of rsw at the inlet and
outlet are greater than the rsw content of the inlet gas, but at some point in the conduit, rsw values go below
this value. To exemplify this situation, three possible paths (linear) of evolution of temperature and pressure
in the conduit are over‐imposed to the color map in Figure 4‐24.
The yellow line represents a situation in which there is a significant pressure drop but a modest drop in
temperature. In this case the rsw will increase along the conduit and free water will therefore never be found
along the conduit.

120
 Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

The red line represents a case in which there is a modest pressure drop but a significant drop in temperature
along the conduit. In this case the rsw will decrease along the conduit and water could drop out at some point
in the conduit. However, in this case, there will be free water at the outlet of the conduit also.
The light blue line represents a case in which there is a significant pressure and temperature drop. In this case
the rsw will first decrease and then increase, exhibiting a local minimum. If, at some point, the value of rsw goes
below the rsw content of the inlet gas, then free water will form.

FIGURE 4‐24. COLOR MAP OF AS A FUNCTION OF PRESSURE AND TEMPERATURE DEPICTING THREE POSSIBLE TRAJECTORIES OF
PRESSURE AND TEMPERATURE IN A CONDUIT (DEPARTING FROM AN INLET PRESSURE AND TEMPERATURE OF 200 BARA AND 70
°C)

121
Fluid Behavior Treatment in Oil and Gas Production Systems M. Stanko

REFERENCES
[4‐1] Whitson, C.H. & Brule, M.R. (2000). Phase Behavior. Society of Petroleum Engineers. 978‐1‐55563‐
087‐4.
[4‐2] Standing, M.B. (1977). Volumetric and Phase Behavior of Oil Field Hydrocarbon Systems. 8th printing,
Society of Petroleum Engineers, Dallas.
[4‐3] Whitson, C. H. (1998). PVT Analysis Manual. Chapter 3. Norsk Hydro.

122
 The Field Development Process M. Stanko

5. THE FIELD DEVELOPMENT PROCESS

The design of an optimum development plan of an offshore hydrocarbon field aims to maximize its economic
value to the stakeholders while producing the resources in a safe and environmentally responsible manner.
This while subjected to variety of socio‐economic, political and regulatory constraints. The challenge is that
most factors contributing to the value of the project are dynamic in nature and are continuously changing over
the lifetime of the field. The evolution and behavior of the physical system (e.g. reservoir and production
system) can be somewhat predicted or controlled but other factors, related to regional and global factors
might change abruptly and unexpectedly as evidenced by historical trends. Some examples of such factors are
cost, consumption, revenue, demands (quantity and quality), political climate and socio‐economic
development.
The field planning process aims to maximize value by performing an educated and robust “guess” of most of
these factors for all available development alternatives. This is done to help taking important decisions that
entail heavy investment and expenditure that must be taken upfront with limited data available (collected
mainly by geophysical and seismic surveys and a few exploration and appraisal wells). The final “truth”
however is always revealed much later, during the execution and operations phase.
Figure 5‐1 shows the typical lifecycle of a hydrocarbon field focusing on the field design phase. Initially, the
value chain model is established, consisting of several critical components that are traditionally considered
and that are of concern for the particular case (to keep simplicity, only a few are shown in the figure). All
components are usually interdependent but the subsurface (reservoir) is central. There are some components
where physical models are defined and used typically to compute their behavior with time with some particular
input (e.g. production profiles). There are other components that require estimating or defining some key
parameters (scheduling, topside structures). There are other components (e.g. economics) where calculations
are performed based on the input from other components and making assumptions of some required factors
(e.g. oil price). A more detailed diagram showing some of the main components considered in the value chain
is shown in Figure 5‐2.
Due to the variety of components considered in the value chain of the project, this usually requires the
involvement and cooperation between several specialists, typically: project management and engineering,
oceanography, marine geo‐technology, engineering geology, marine structures, pipeline engineering, marine
operations, subsea technology, subsea facilities, process technology, top‐side facilities engineering, technical
safety, cost engineering, geography and impacts analyst (environmental, socio‐economical, impact on fishery),
field architecture, etc.

123
 The Field Development Process M. Stanko

FIGURE 5‐1. FIELD DEVELOPMENT TIMELINE AND THE EVOLUTION OF THE VALUE CHAIN MODEL AFTER DECISION ARE MADE

FIGURE 5‐2. DETAILED VALUE CHAIN COMPONENTS

During early phases of field planning some components will have several possible alternatives (e.g. offshore
structures, scheduling) that in turn affect other components. Additionally, most parameters will have an
associated uncertainty (that is often described statistically). With the value chain model, it is possible to
establish all development options and further calculate their associated economic indicators, impact and risks.
The field design process progresses by gradually discarding non‐attractive alternatives and narrowing further

124
 The Field Development Process M. Stanko

the alternatives, factors and details in each individual component. This is typically done through decision gates
(DG).
The field design process aims to find an optimum balance between flexibility and cost for the particular asset
under study. High flexibility is desirable to cope with future changes, field expansion, market fluctuations;
however, it usually comes at a very high cost. For example, oversizing the processing facilities to allow
production ramp‐ups entrains big investments that would likely affect negatively the net present value of the
project. Low flexibility gives less costs but it makes the system very rigid to absorb future changes. The
optimum lies someplace in between.
The decision‐making process within field design should be done leaving an appropriate amount of flexibility
and options open in each stage. This to allow adapting to new information gathered at a later stage and have
the possibility to execute changes when necessary. It also should carry further all relevant uncertainties that
could impact the value of the project.
During the design process the company has the crucial role to look at the solutions proposed by the vendor,
verify their purpose and determine their relevance and applicability for the particular case. Pre‐packaged
solutions that have high flexibility and multiple components are easier to handle from the contracting point of
view but they might cause extra expenses that affect negatively the economic value of the asset. Strong
cooperation between company and vendor and performing third party “fit for purpose” reviews are ways to
ensure that the solutions offered are applicable and necessary for the particular field development.
Oil and gas companies usually have an internal project development process similar to the one shown in Figure
5‐3. Along the process there are decision gates (DG, usually located between phases) where the status of the
project is reviewed and decisions are made to continue, review or terminate it. A brief description of each
phase is given next.

Discovery DG0 DG1 DG2 DG3

Pre‐ Exploration Feasibility Concept Pre

exploration and appraisal studies planning engineering

Identification of Project planning

Statement of
business case PDO
Commerciality DG4
(SOC)

Abandonment
Detailed Construction Testing and Operations and
Engineering start‐up decommissioning

Project execution

FIGURE 5‐3. FIELD DEVELOPMENT PROCESS

125
 The Field Development Process M. Stanko

5.1. BUSINESS CASE IDENTIFICATION

The main goal of this step is to prove economic potential of the discovery and quantify and reduce the
uncertainty in the estimation of reserves.
It usually involves the following steps:
 Pre‐exploration – scouting: collecting information on areas of interests. In this step technical, political,
geological, geographical, social, environmental considerations are taken into account. E.g. expected
size of reserves, political regime, government stability, technical challenges of the area, taxation
regime, personnel security, environmental sensitivity, previous experience in the region, etc.
 Getting pre‐exploration access – The exploration license (usually non‐exclusive). In the NCS only
seismic and shallow wells are allowed. This is usually done by specialized companies selling data to oil
companies.
 Identify prospects.
 Apply and obtain exclusive production license. In the NCS18: Licensing rounds (frontier areas) or
Awards in predefined areas (APA). The current fees are 34.000 NOK/km2 for the first year, 68.000
NOK/km2 for the second year and 137.000 NOK/km2 per year thereafter.
 Exploration. Perform geological studies, geophysical surveys, seismic, exploration drilling (Well cores,
wall cores, cuttings samples, fluid samples, wireline logs, productivity test).
 Discovery.
 Assessment of the discovery and the associated uncertainty. Risk management.
 Probabilistic reserve estimation. Identify and assess additional segments.
o Perform simplified economic valuation of the resources.
o Field appraisal to reduce uncertainty: more exploration wells and seismic to determine for
example: fault communication, reservoir extent, aquifer behavior, location of water oil
contact or gas oil contact.
Possible outcomes of DG0 are:
 Issue a SOC (Statement of Commerciality) and proceed with development.
 Continue with more appraisal
 Sell the discovery.
 Do nothing (wait)
 Relinquish to the government
5.1.1. RESERVE ESTIMATION USING PROBABILISTIC ANALYSIS
A typical problem in field development is estimating initial hydrocarbon in place. For example, consider the
simple expression to compute initial oil in place (N) of a clean (no shale) undersaturated oil layer:

𝑉 ∙𝜙∙ 1 𝑆 EQ. 5‐1

𝑁
𝐵
Where:
𝑉 Rock volume (in m3)
𝜙 Porosity (fraction)
𝑆 Water saturation (fraction)

18
NCS. Norwegian Continental Shelf

126
 The Field Development Process M. Stanko

𝐵 Oil formation volume fraction

The input to this equation is often not a unique set of values but rather a range. This could be due to
uncertainty: e.g. lack of more detailed information, errors in measurement, or simply due to natural variability
of the parameters. Additionally, often within this range there are some values that have a higher probability
of occurrence than others. Therefore, input is typically characterized with a probability distribution defined
between a lower and upper limit and that provides a probability for values of the variable. Examples of
distributions are uniform (all values have the same probability), normal, triangular (both exhibit a peak).
The variability in the input often causes the output to be variable, and, a priori, uncertain (rather than a unique
value, when the input are single numbers). This is shown in Figure 5‐4. For our example on initial oil in place
estimation, Y is N and X1, X2, X3 and X4 are porosity, rock volume, water saturation and oil formation volume
factor. The “model” is simply Eq. 5‐1. However, the ellipse in Figure 5‐4 can also represent the result of solving
a system of equations, a simulator, a process, etc.

FIGURE 5‐4. MODEL OR SIMULATION WITH UNCERTAINTY IN ITS INPUT PARAMETERS

To quantify and analyze the uncertainty in the output one can employ sampling methods. The goal of sampling
methods is to compute several values that are the “most representative” of the function of interest and
perform a frequency analysis on them to compute its probability distribution. Even though it is impossible to
sample all possible values of the function (this will require an infinite number of samples to map thoroughly
the output domain), if the number of samples taken is high enough (or properly distributed), it should be a
good estimate of the “real” probability distribution of the function.
With the probability distribution of the function, one can then estimate the mean, the expected “spread” of
the distribution and other useful quantities such as percentiles (the most common are P10, P50 and P90). The
percentiles are found from the cumulative probability plot by intersecting the percent value with the curve
and reading the value on the x axis. For example, if the cumulative probability plot was found by compounding
probabilities from largest to lowest value, a value of P50 means that there is 50% probability that the variable
is equal or greater than P50. On the contrary, if the cumulative probability curve was found by compounding
probabilities from smallest to largest, a value of P50 means that there is 50% probability that the variables is
equal or smaller than P50.
Sampling is typically performed by generating sets of input variables that represent the variability of the input.
These sets are further input to the model individually, the model is executed (this is often referred to as a

127
 The Field Development Process M. Stanko

“simulation”) and all outputs are recorded. A frequency analysis is then executed on the output to find the
probability distribution.
One popular sampling method is Monte Carlo. This method consists of choosing randomly a value of the
cumulative distribution function (cdf) of each input variable (if using a continuous probability function), obtain
a value of the variable and combine it with random values of other variables. It is advantageous to sample on
the cumulative distribution function as the number is bounded between 0 and 1. It also guarantees the
resulting sample of the input exhibits the same trend as the original probability distribution.
As an example, Figure 5‐5 shows the result of a Monte Carlo simulation19 of Eq. 5‐1, with increasing number
of samples. Above 1000 samples, the frequency distribution changes very little. By increasing the number of
samples, the probability distribution converges to a unique (the real) distribution.

101 samples 102 samples 103 samples

104 samples 105 samples 106 samples

FIGURE 5‐5. PROBABILITY DISTRIBUTION OF INITIAL OIL IN PLACE CALCULATED WITH MONTE CARLO SIMULATION AND DIFFERENT
NUMBER OF SAMPLES

To estimate the required number of iterations to perform in the Monte Carlo method, one uses statistical
inference. Consider that the probability distribution function of N is known and it displays a normal probability
distribution with mean 𝜇. If a sample of size “n” is taken from that “real” distribution (this is what Monte Carlo
does), the quantity T.

𝑋 𝜇 EQ. 5‐2
𝑇
𝑆
√𝑛
will be t‐distributed symmetrically around zero, with n‐1 degrees of freedom. 𝑋 is the mean of the sample and
S is the standard deviation of the sample.

19
Using uniform distribution for porosity (0.18‐0.3), rock volume (2000‐2500 MMbbl), oil saturation (0.8‐0.9) and oil
formation volume factor (1.35‐1.6)

128
 The Field Development Process M. Stanko

One can define an “interval of confidence” for T (Eq. 5‐3). For example, if 25 samples were taken (24 degrees
of freedom) and A is 1.96, this means that there is a 95% probability that T is located within the interval. This
is often referred to as a 95% confidence interval. The higher the confidence level, the larger A must be.
EQ. 5‐3
𝐴 𝑇 𝐴

The value A (often referred to as t, with α in this case being 100‐95=5) is affected by the number of samples
and by the confidence level, and it can be computed from the cdf of the t‐distribution A. However, for a large
number of samples (e.g. more than 30), the t‐distribution tends to overlap with a normal distribution with zero
mean and a standard deviation of one.
Substituting the expression of T in Eq. 5‐3, and rearranging terms:

𝑆 𝑆 EQ. 5‐4
𝑋 𝑡 / ∙ 𝜇 𝑋 𝑡 / ∙
√𝑛 √𝑛
Eq. 5‐4 then defines a range for the real (unknown) mean 𝜇 if the confidence interval (A) is provided.
As an example, consider a normal distribution of N, with real (unknown) mean of 660 MMstb and standard
deviation of 150 stb. Figure 5‐6 shows the mean and confidence interval (for a confidence level of 95 %) of a
sample of varying sizes. The confidence interval gets smaller and smaller as the number of samples is increased
and the distribution of the sample resembles closer the “real” distribution.

7 samples 37 samples

133 samples 517 samples

FIGURE 5‐6. PROBABILITY DISTRIBUTION OF INITIAL OIL IN PLACE SAMPLED FROM A NORMAL DISTRIBUTION FOR DIFFERENT
NUMBER OF SAMPLES

129
 The Field Development Process M. Stanko

To estimate the number of samples required in Monte Carlo, one can modify Eq. 5‐4 by specifying the error
(e) one wishes to achieve.

𝑆 EQ. 5‐5
𝑒 𝑡 / ∙
√𝑛
e is the absolute error, and it is often expressed in terms of the mean of the sample, 𝑋, for example, by
multiplying it with a fraction F (between 0‐1). By substituting this in Eq. 5‐5, and clearing “n” it yields

𝑆 EQ. 5‐6
𝑛 𝑡 / ∙
𝐹∙𝑋
To estimate a value of n, one must have performed at least one iteration, and S and 𝑋 should be available.
There will be a new value of n for each iteration and one stops when n estimated is smaller than the actual
sample size.
Sometimes the number of iterations required for the Monte Carlo method to achieve an acceptable error are
too large to be practical, especially if the simulation running time is high. There are other methods that
perform a smarter sampling of the function (instead of random), that require less iterations. One of this
methods is Latin hypercube sampling. In Latin Hypercube sampling the steps are:
 Define the number of samples “n” to use.
 Subdivide the pdf of each input variable in “n” intervals of equal probability.
 Create a set of “n” values for each variable by picking a random number in each interval (using the cdf
of each interval, just like in Monte Carlo).
 Create “n” input sets by picking randomly a value of each variable set. It is not allowed to pick the
same value twice.
 Run the simulations with the “n” input sets and apply a frequency analysis to the results.

5.2. PROJECT PLANNING

The main goal of the planning phase is to perform a systematic screening of concepts, to define a preferred
development concept and to evaluate its profitability, technical feasibility and HSE within acceptable levels of
uncertainty. Furthermore, to document the solution for delivery to the authorities managing the production
license.

5.2.1. FEASIBILITY STUDIES

The main goal of this step is to justify further development of the project, finding one or more concepts that
are technically, commercially and organizationally feasible. Some specific tasks of this phase are:

 Define objectives of the development in line with the corporate strategy.

 Establish feasible development scenarios.
 Create a project timeline and a workplan.
 Identify possible technology gaps and blockers.
 Identify the needs for new technology.
 Identify added value opportunities.
 Cost evaluation for all options.
5.2.2. CONCEPT PLANNING (LEADING TO DG2)
Identify development concepts, rank them and select and document a viable concept (Base Case Scenario).

 Evaluate and compare alternatives for development and screen out non‐viable options.

130
 The Field Development Process M. Stanko

 Elaborate a Project Execution Plan (PEP) which describes the project and management system.
 Define the commercial aspects, legislation, agreements, licensing, financing, marketing and supply,
taxes.
 Create and refine a static and a dynamic model of reservoir. Define the depletion and production
strategy.
 Define an HSE program.
 Flow assurance evaluation. Identification of challenged related with fluid properties, multiphase
handling and driving pressure.
 Drilling and well planning.
 Pre‐design of facilities.
 Planning of operations, start‐up and maintenance.
 Cost and manpower estimates of the best viable concept.
5.2.3. FIELD PRODUCTION PROFILE AND ECONOMIC VALUE
During the concept phase, one of the main tasks is to define the field production schedule that provides the
maximum economic value for the project. The economic value of the project could be estimated using
different financial evaluation approaches; one of the approaches is the net present value (NPV), calculated on
a yearly basis, which is defined as:

𝐶𝐹
𝑁𝑃𝑉 EQ. 5‐7
1 𝑖

Where:
𝑖 Discount rate (usually a value in the range 5‐12%)
𝑡 Integer counter for the number of years
𝑛 Total number of years
𝐶𝐹 Cash flow of year “t”

Neglecting royalties and tax payments (which vary from country to country, or even on license type), the
cashflow for the project can be expressed as:
EQ. 5‐8
𝐶𝐹 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑂𝑃𝐸𝑋 𝐷𝑅𝐼𝐿𝐿𝐸𝑋 𝐶𝐴𝑃𝐸𝑋

During the first years of the project, when the field is under construction and there is no production, only
drilling expenditures (also known as DRILLEX) and capital expenditures (CAPEX) are considered. CAPEX are
expenditures to acquire, design, manufacture and transport physical assets such as an offshore structure,
topside facilities, subsea system, etc. For later years, when the field is producing only revenue and operating
expenditures (OPEX) are considered.
The CAPEX related with the offshore structure and topside facilities depends strongly on the type and weight
of units and equipment that will be placed on the offshore structure. The type of units and equipment depend
on the treatment processes required for reservoir fluids, while the weight is given mainly by the maximum
liquid and gas processing capacities. Therefore, the offshore structure and topside CAPEX is a function of the
maximum field liquid rate (ql,max) and the maximum field gas rate (qg,max) produced during the life of the field.
The relationship is often assumed linear, as shown in equation Eq. 5‐9.
EQ. 5‐9
𝐶𝐴𝑃𝐸𝑋 𝐶 ∙𝑞, 𝐶 ∙𝑞 , 𝐶

131
 The Field Development Process M. Stanko

For some offshore structures that “house” wells (for example jackets, gravity‐based platforms, tension leg
platforms with dry Christmas trees), the CAPEX of offshore and topside also depends strongly of the number
of wells.
The subsea system CAPEX depends of several factors: number of wells, number of subsea manifolds and
flowlines, umbilicals, water depth, etc. Despite showing dependencies on these many factors, the subsea
CAPEX can be often expressed as function (often linear) with the number of wells.
EQ. 5‐10
𝐶𝐴𝑃𝐸𝑋 𝐶 𝐶 ∙𝑁
Where:
𝑁 Number of wells

DRILLEX is typically computed the well cost times the number of wells:
EQ. 5‐11
𝐷𝑅𝐼𝐿𝐿𝐸𝑋 𝑁 ∙𝑃
Where:
𝑃 Cost of drilling a well

Yearly OPEX depends on the producing rates of oil, gas and water, and the number of wells in operation in
year “t”. Additionally, there are other yearly costs such as energy usage, insurance, transportation of goods
and personnel to and from the field, consumption of chemicals, among others. The relationship is often
assumed linear:
EQ. 5‐12
𝑂𝑃𝐸𝑋 𝑂 ∙𝑞, 𝑂 ∙𝑞 , 𝑂 ∙𝑁 , 𝑂

The revenue function will be discussed next with an example. Due to the discounting factor, the production in
later years typically contributes less to the NPV than production in earlier years. Thus, one usually tends to
favor production in earlier years against production in later years.
Estimating the NPV of the revenue, case study: oil field with a linear production potential curve, operating in
plateau mode
Assume the field exhibits the following dimensionless production potential linear equation for oil production:

𝑁𝑝 EQ. 5‐13
𝑞𝑝𝑝 1 𝛼∙
𝑁

The production potential is then:

𝑁𝑝 EQ. 5‐14
𝑞𝑝𝑝 𝑞𝑝𝑝𝑜,𝑓𝑖𝑒𝑙𝑑 ∙ 1 𝛼∙
𝑁

If there are several identical producing wells in the production systems that are independent from each other,
then 𝑞𝑝𝑝𝑜,𝑓𝑖𝑒𝑙𝑑 is:

𝑞 𝑁 ∙𝑞 EQ. 5‐15
, ,

If the field is produced at plateau and then decline, the plateau duration (in years) can be calculated with the
following equation:

132
 The Field Development Process M. Stanko

𝑁 1 1 EQ. 5‐16
∆𝑡 ∙
𝛼∙𝑡 𝑞 , 𝑞 ,

To derive an expression of field production rate versus time, we follow a similar procedure to the one
presented in Example 1 of section 1.6, assuming the field is produced in plateau mode and then declines, and
that production starts at 𝑡𝑖𝑛𝑖 , instead of at zero the field rate is given by:
EQ. 5‐17
𝑓𝑜𝑟 𝑡 𝑡 𝑞𝑓 0

EQ. 5‐18
𝑓𝑜𝑟 𝑡 𝑡 𝑡 ∆𝑡 𝑞 𝑞 ,

∙ ∙ EQ. 5‐19
𝑓𝑜𝑟 𝑡 𝑡 ∆𝑡 ,
∙ ∆
𝑞 𝑞 , ∙𝑒

Here t is input in years.

The recovery factor can be computed with:

∆
1 ∙ ∆ EQ. 5‐20
𝑅 𝑞 , ∙𝑡 ∙ 𝑑𝑡 𝑞 , ∙𝑡 ∙𝑒 𝑑𝑡
𝑁 ∆

Which gives the following expression:

𝑞 , ∙𝑡 1 ∙ ∆ EQ. 5‐21
𝑅 ∙ ∆𝑡 ∙ 1 𝑒
𝑁 𝑚

Returning to the expression of the project net present value:

𝐶𝐹 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑂𝑃𝐸𝑋 𝐷𝑅𝐼𝐿𝐿𝐸𝑋 𝐶𝐴𝑃𝐸𝑋 EQ. 5‐22

𝑁𝑃𝑉
1 𝑖 1 𝑖

If we separate each term in the sum:

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑂𝑃𝐸𝑋 𝐷𝑅𝐼𝐿𝐿𝐸𝑋 𝐶𝐴𝑃𝐸𝑋 EQ. 5‐23

𝑁𝑃𝑉
1 𝑖 1 𝑖 1 𝑖 1 𝑖

Each term is renamed as follows:

𝑁𝑃𝑉 𝑁𝑃𝑉 𝑁𝑃𝑉 𝑁𝑃𝑉 𝑁𝑃𝑉 EQ. 5‐24

To find an analytical equation for the 𝑁𝑃𝑉 , we will use continuous discounting instead of yearly discounting.
The discrete discount factor (DF) is:

𝐷𝐹 𝑡 1 𝑖 EQ. 5‐25

If discounting is made over “x” discrete periods within the year, the expression must be modified as follows:

∙
𝑖 EQ. 5‐26
𝐷𝐹 𝑡 1
𝑥

133
 The Field Development Process M. Stanko

The limit when x is large yields:

𝑥∙𝑡
𝑖 EQ. 5‐27
lim 𝐷𝐹 𝑡 lim 1 𝑒𝑖∙𝑡
𝑥→∞ 𝑥→∞ 𝑥

Assuming that there are sales of both oil and gas, at a constant price of 𝑃𝑜 , 𝑃𝑔 respectively, assuming the field
will produce a constant gas‐oil ratio (𝑅𝑝 ) and using a continuous discounting with the exponential function
and discount rate “𝑖” the net present value of the revenue stream at a given time “t” is calculated by:

𝑁𝑃𝑉 𝑞 𝑡 ∙𝑡 ∙ 𝑃 𝑅 ∙𝑃 ∙𝑒 ∙
𝑑𝑡 EQ. 5‐28

Expanding the expression for the plateau and post‐plateau periods:

∆
∙
𝑁𝑃𝑉 𝑞 , ∙𝑡 ∙ 𝑃 𝑅 ∙𝑃 ∙𝑒 𝑑𝑡
EQ. 5‐29
∙ ∆ ∙
𝑞 , ∙𝑡 ∙𝑒 ∙ 𝑃 𝑅 ∙𝑃 ∙𝑒 𝑑𝑡
∆

Here, the following nomenclature has been adopted:

𝛼∙𝑞 , ∙𝑡 EQ. 5‐30

𝑚
𝑁

Solving the integral for the plateau period:

∙ ∙ ∆
𝑒 𝑒
𝑁𝑃𝑉 𝑞 , ∙ 𝑃 𝑅 ∙𝑃 ∙𝑡 ∙ 𝑞 , ∙ 𝑃 𝑅 ∙𝑃
𝑖 EQ. 5‐31
∙ ∆ ∙
∙𝑡 ∙𝑒 ∙ 𝑒 𝑑𝑡
∆

Solving the integral for the decline period:

∙ ∙ ∆
𝑒 𝑒
𝑁𝑃𝑉𝑟𝑒𝑣 𝑞 , ∙ 𝑃 𝑅 ∙𝑃 ∙𝑡 ∙ 𝑞𝑝,𝑓 ∙ 𝑃 𝑅 ∙𝑃
𝑖 EQ. 5‐32
𝑚 𝑖 ∙ ∆ 𝑚 𝑖 ∙𝑡
∙ ∆
𝑒 𝑒
∙𝑡 ∙𝑒 ∙
𝑚 𝑖

Rearranging terms:

𝑁𝑃𝑉 𝑞 , ∙ 𝑃 𝑅 ∙𝑃 ∙𝑡
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆ EQ. 5‐33
𝑒 𝑒 𝑒 𝑒
∙
𝑖 𝑚 𝑖

To study the behavior of 𝑁𝑃𝑉 , the following values are used (some taken from Nunes[5‐1]):

 𝑡 = 25 years

134
 The Field Development Process M. Stanko

 N = 2.19𝐸09 𝑠𝑡𝑏
 qppo,well = 20 000 stb/d
 i = 0.09
 α=4
 𝑃 50 𝑈𝑆𝐷/𝑠𝑡𝑏
 𝑃 0.004 𝑈𝑆𝐷/𝑠𝑐𝑓
 Rp = 200 scf/stb
 𝑡 352 𝑑/𝑦𝑒𝑎𝑟
 𝑡 = 5 years
NPVrev is calculated until abandonment time, which is:
EQ. 5‐34
𝑡 𝑡 𝑡

Using these values, the calculation of the revenue net present value is depicted in Figure 5‐7 for two (2)
number of producing wells and several values of oil field plateau rate. The function has been plotted up to the
maximum oil plateau rate that it is physically possible to produce from the field (e.g. 240 000 for 12 wells and
300 000 for 15 wells respectively).

FIGURE 5‐7. BEHAVIOR OF THE NET PRESENT VALUE OF THE REVENUE VERSUS OIL PLATEAU RATE FOR TWO NUMBERS OF WELLS

The function “growth” slows down when the oil plateau rates approaches its upper bound. This indicates that
the production strategy to produce as much as possible as early as possible indeed does increase the revenue
present value, but its effect becomes weaker and weaker as the rate approaches the maximum rate the field
can produce.
The behavior of NPV, i.e. when combining revenues and expenses in a cash flow and discounting in time will
be discussed next with an example.
Estimating the NPV, case study: offshore oil field with a linear production potential curve, operating in plateau
mode
In this case we will use some data presented by Nunes[5‐1] . We make the following assumptions:

 CAPEX and DRILLEX are paid at initial time, thus it is not necessary to discount them.
 The maximum liquid capacity is taken as equal to the oil plateau rate, i.e., no water will be produced
from the field.

135
 The Field Development Process M. Stanko

 Each producer well has an associated water injector, thus the total number of wells 𝑁 is equal to
2∙𝑁 ,
 All wells are drilled before production starts.
The capital expenditures of the topside facilities and the offshore structure can be estimated with the following
expression, using the oil plateau rate (maximum oil rate expected), and assuming that the producing gas‐oil
ratio, Rp is constant:
EQ. 5‐35
𝐶𝐴𝑃𝐸𝑋 𝐶 ∙𝑞 , 𝐶 ∙𝑞 , , ∙𝑅 𝐶 𝐶 ∙𝑞 , 𝐶 ∙𝑞 , ∙𝑅 𝐶

With:
 𝐶 =1.2 E03 USD/stb/d
 𝐶 =6.6 USD/scf/d
 𝐶 =1.07 E09 USD
The capital expenditures of the subsea system can be estimated with the following expression. The last term
is the cost of a subsea manifold assuming that there is a maximum of 4 wells per manifold.

𝑁 EQ. 5‐36
𝐶𝐴𝑃𝐸𝑋 𝐶 𝐶 ∙𝑁 𝐶 ∙ 𝑟𝑜𝑢𝑛𝑑𝑢𝑝
4
 𝐶 =130 E06 USD
 𝐶 =66.1 E06 USD/well
 𝐶 =32 E06 USD/manifold
The cost of each well is 𝑃 =150 E06 USD/well
OPEX is varying in time, as a function of the oil and gas rate in time. The following expression will be assumed:

𝑂𝑃𝐸𝑋 𝑡 𝑂 ∙𝑞 𝑡 𝑂 ∙𝑞 𝑡 𝑂 ∙𝑁 𝑂 EQ. 5‐37

𝑂 ∙𝑞 𝑡 𝑂 ∙𝑞 𝑡 ∙𝑅 𝑂 ∙𝑁 𝑂

With:
 𝑂 =400 USD/stb/d
 𝑂 =0 USD/scf/d
 𝑂 =0.7 E06 USD/well
 𝑂 =80 E06 USD
OPEX will be discounted continuously, in a similar manner as the revenue:

𝑁𝑃𝑉 𝑂𝑃𝐸𝑋 𝑡 ∙ 𝑒 ∙
𝑑𝑡 EQ. 5‐38

The derivation gives:

𝑁𝑃𝑉 𝑞𝑝,𝑓 ∙ 𝑂 𝑂 ∙𝑅
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆
𝑒 𝑒 𝑒 𝑒
∙ EQ. 5‐39
𝑖 𝑚 𝑖
∙ ∙
𝑒 𝑒
𝑂 ∙𝑁 𝑂
𝑖

136
 The Field Development Process M. Stanko

The expression is similar to the one obtained or NPVREV, thus it is possible to combine both into one:

𝑁𝑃𝑉 𝑞𝑝,𝑓 ∙ 𝑃 𝑅 ∙𝑃 ∙𝑡 𝑂 𝑂 ∙𝑅
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆
𝑒 𝑒 𝑒 𝑒
∙ EQ. 5‐40
𝑖 𝑚 𝑖
∙ ∙
𝑒 𝑒
𝑂 ∙𝑁 𝑂
𝑖

Figure 5‐8 shows, for 12 producers (24 wells in total) the NPVrev, NPVOPEX, the project NPV, and the
CAPEX+DRILLEX versus the oil field plateau rate.

FIGURE 5‐8. PROJECT NPV, NPVREV , NPVOPEX, AND CAPEX+DRILLEX FOR 12 PRODUCER WELLS VERSUS OIL PLATEAU RATE.

The NPV curve shows a maximum at around 220 000 stb/d. This maximum corresponds to the point where:

𝜕𝑁𝑃𝑉 EQ. 5‐41

0
𝜕𝑞 ,

As mentioned earlier, the NPV is a function of revenue, CAPEX, OPEX, therefore:

𝜕𝑁𝑃𝑉 𝜕𝐶𝐴𝑃𝐸𝑋 𝜕𝐷𝑅𝐼𝐿𝐿𝐸𝑋 𝜕𝑁𝑃𝑉 EQ. 5‐42

0
𝜕𝑞 , 𝜕𝑞 , 𝜕𝑞 , 𝜕𝑞 ,

As we have assumed the CAPEX is a linear function of the plateau oil production, the maximum of NPV is given
when:

𝜕𝑁𝑃𝑉 EQ. 5‐43

𝐶 𝐶 ∙𝑅
𝜕𝑞 ,

Effect of water production

In practice, the field will also produce water. Considering this, it is possible to express the CAPEX as a function
of both water and hydrocarbon flowrates using the water cut (WC), as shown in Eq. 5‐44

137
 The Field Development Process M. Stanko

𝑞 , EQ. 5‐44
𝐶𝐴𝑃𝐸𝑋 𝐶 ∙ 𝐶 ∙𝑞 , ∙𝑅 𝐶
1 𝑊
The maximum liquid and gas rates often occur during the plateau period, therefore:

𝑞 , EQ. 5‐45
𝐶𝐴𝑃𝐸𝑋 𝐶 ∙ 𝐶 ∙ 𝑞, ∙ 𝑅 𝐶
1 𝑊
To calculate the slope of the CAPEX function, Eq. 5‐45 is differentiated once.

𝜕𝐶𝐴𝑃𝐸𝑋 𝐶 EQ. 5‐46

𝐶 ∙𝑅
𝜕𝑞 , 1 𝑊

This function will give another maximum and optimum oil plateau rate of the NPV function compared with the
one given in Eq. 5‐9. The CAPEX slope will be higher, thus the optimum plateau rate will be lower than for the
case without water production.
Effect of number of producers
Figure 5‐9 shows NPV values calculated for number of producer wells between 1 and 20 and oil plateau rates
between 5 000 and 400 000 stb/d. The white space are plateau rates that are not feasible because they exceed
the maximum production possible by the number of producers specified. The NPV exhibits a global maximum
indicated with the blue dot (9 producers, oil plateau rate 170 186 stb/d). The black line depicts combinations
that give a NPV of zero.

FIGURE 5‐9. COLOR CONTOUR OF NPV VERSUS NUMBER OF PRODUCING WELLS AND FIELD PLATEAU RATE

Effect of initial oil in place

Figure 5‐10 shows NPV values calculated for number of producer wells between 1 and 20 and oil plateau rates
between 5 000 and 400 000 stb/d, for three different values of initial oil in place (base case, 40% less and 40%
more). The white space are plateau rates that are not feasible because they exceed the maximum production

138
 The Field Development Process M. Stanko

possible by the number of producers specified. The NPV exhibits a global maximum indicated with the blue
dots:
 An N value 40% smaller than the base gives an optimum at 5 producers, oil plateau rate 95 127 stb/d.
 An N value 40% bigger than the base gives an optimum at 12 producers, oil plateau rate 227 814
stb/d.

a) N= 0.6·base b) N= base c) N= 1.4·base

FIGURE 5‐10. COLOR MAP OF NPV VERSUS NUMBER OF PRODUCING WELLS AND FIELD PLATEAU RATE, FOR THREE VALUES OF
INITIAL OIL IN PLACE. THE BLUE DOT INDICATES THE OPTIMAL COMBINATION OF NUMBER OF PRODUCERS AND OIL PLATEAU RATE

Effect of oil price

Figure 5‐11 shows NPV values calculated for number of producer wells between 1 and 20 and oil plateau rates
between 5 000 and 400 000 stb/d, for two values of oil price, 50 and 70 USD/stb. The white space are plateau
rates that are not feasible because they exceed the maximum production possible by the number of producers
specified. The NPV for the oil price of 70 USD/stb exhibits a global maximum at 12 producers, oil plateau rate
225 291 stb/d.

a) PO= 50 USD/stb b) PO= 70 USD/stb

FIGURE 5‐11. COLOR MAP OF NPV VERSUS NUMBER OF PRODUCING WELLS AND FIELD PLATEAU RATE, FOR OIL PRICE 50
USD/STB AND 70 USD/STB.

139
 The Field Development Process M. Stanko

5.2.4. PRE‐ENGINEERING (LEADING TO DG3)

Further mature, define and document the development solution based on the selected concept. Some specific
tasks are:

 Selection of the final technical solution. Decide and define all remaining critical technical alternatives.
 Execute Front‐End Engineering Design (FEED): determine technical requirements (arranged in
packages) for the project based on the final solution chosen. Estimate cost of each package.
 Plan and prepare the execution phase.
 Prepare for submission of the application to the authorities.
 Perform the Environmental impact assessment.
 Establish the basis for awarding contracts.
The outcome of DG3 is[5‐2]:
 Issue plan for development and operations (PDO), plan for Installation and Operation of Facilities for
transport and utilization of petroleum (PIO), and Impact assessment report.

5.3. PROJECT EXECUTION

If the plan for development and operations is approved by the authorities, the project execution phase starts.

5.3.1. DETAILED ENGINEERING, CONSTRUCTION, TESTING AND STARTUP

 Detailed design, procurement of the construction materials, construction, installation and
commissioning of the agreed facilities. This can be done in two ways:
o Individual contracts
o Detailed engineering
o Bids, contracts
o Construction, fabrication
o Installation
o Commissioning (Cold or Hot)
o Or using an EPCM (Engineering, procurement, construction, and management contract) with
one main contractor.
 Constructing wells.
 Perform hand over to asset, operations
 Prepare for start‐up, operation and maintenance

5.4. OPERATIONS

 Production startup, Build‐up phase, Plateau phase, decline phase, Tail production, Field shut‐down.
 Maintenance.
 Planning Improved Oil recovery methods.
 Allocation and metering.
 De‐bottlenecking.
 Troubleshooting.

5.5. DECOMMISSIONING AND ABANDONMENT

 Engineering “down and clean”: flushing and cleaning tanks, processing equipment, piping.
 Coordinate with relevant environmental and governmental authorities.
 Well plugging and abandonment (P&A)
 Cut and remove well conductor and casing.
 Remove topside equipment.

140
The Field Development Process M. Stanko

 Removal of the offshore structure: Lifting operations and transport

 Remove or bury subsea pipelines
 Mark and register leftover installations on marine maps
 Monitoring
 Recovery of material: Scrap (steel) and recycling equipment (Gas turbines, separators, heat
exchangers, pumps, processing equipment)
 Disposal of residues

141
The Field Development Process M. Stanko

REFERENCES
[5‐1] Nunes, G. C.; Da Silva, A. H & Esch, L.G. (2018). A Cost Reduction Methodology for Offshore Projects.
OTC‐28898‐MS. Offshore Technology Conference. Houston.
[5‐2] Norwegian Petroleum Directorate. Guidelines for PDO and PIO. Retrieved from
http://www.npd.no/Global/Engelsk/5‐Rules‐and‐regulations/Guidelines/PDO‐PIO‐
guidelines_2010.pdf, on Jan 9th, 2017.
[5‐3] Presentation “Field Development and Portfolio Evaluation”. Statoil. Retrieved from
http://www.uio.no/studier/emner/matnat/math/MEK4450/h15/ppt/l1‐2/10‐field‐development‐
and‐cvp‐process‐august‐2015.pdf on Jan 9th, 2017.
[5‐4] Jahn, F., Cook, M. & Graham, M. (2008). Hydrocarbon Exploration and Production. 2nd edition.
Elsevier Science. 978‐0‐08056‐8836.

142
 Offshore Structures for Oil and Gas Production M. Stanko

6. OFFSHORE STRUCTURES FOR OIL AND GAS PRODUCTION

In this section a brief discussion will be made about offshore structures typically used for oil and gas
production. Some particular offshore vessels and structures that are not discussed here are drilling vessels and
platforms, well intervention vessels, vessels used to transport and lay down pipelines and equipment, supply
vessels and tankers. Some fields can also be developed subsea and their production tied to shore (subsea‐to‐
beach) or to existing installations. These cases are not discussed here. The current records of subsea tiebacks
are the Penguin A‐E field for oil (69.8 km) in the North Sea and the Tamar field offshore Israel (149.7 km).
Offshore structures for oil and gas production have, in general, some of the components provided in the list
below:
 Facilities for drilling and full intervention. This includes drilling tower, BOP, drilling floor, mud package,
cementing pumps, storage deck for drill pipes and tubulars, drilling risers.
 Facilities for light well intervention.
 Processing facilities: separator trains for primary oil, gas and water separation, gas processing train,
water processing train.
 Gas injection system
 Gas compression units for pipeline transport
 Water injection system
 Living quarters
 Helideck.
 Power generation.
 Flare system.
 Utilities (hydraulic power fluid, compressed air, drinking water unit, air condition system, ventilation
and heating system)
 Bay for wellheads and christmas trees
 Production manifolds
 Oil storage
 Facilities for oil offloading
 Control system
 Monitoring system
 System for storage, injection and recovery of production chemicals (wax, scale, hydrate or corrosion
inhibitors)
 Repair workshop
Not every offshore production structure has all elements mentioned on the list. The required functionality will
vary depending on the type and processing capacity required for reservoir fluids, number of wells required,
the development plan and future modifications to be made, the architecture of the production system, among
others. It is also not uncommon to have two structures or more in the field with complementary functionality.
Figure 6‐1 shows some common types of marine structures that are typically used for offshore oil and gas
exploitation classified under two main categories: Bottom‐supported or floating.

143
 Offshore Structures for Oil and Gas Production M. Stanko

Fixed Compliant

BOTTOM‐SUPPORTED

Gravity‐Based
Jacket Jack‐up Compliant tower
Structure
Positively
Neutrally buoyant
buoyant
FLOATING

Tension Leg
Ship FPSO Semi‐Sub Sevan FPSO Spar
Platform (TLP)
FIGURE 6‐1. SOME COMMON MARINE STRUCTURES FOR OIL AND GAS EXPLOITATION

Bottom‐supported structures display reduced movement in the lateral and vertical direction. As the name
indicates, most of the weight and the environmental loads on the structure are transferred to the seabed soil.
Compliant towers have some lateral movement because they are allowed to rotate about their base.
Floating structures keep above the water level due to buoyancy and have relevant movement in the lateral
and vertical direction due to environmental loads such as wind, current and waves. They are commonly
moored to avoid drifting excessively with free hanging lines (steel catenary, Figure 6‐2a), with pre‐tensioned
lines (taut, Figure 6‐2b) or a combination of both. The buoyancy is controlled actively with ballast depending
on the amount of fluids stored onboard.

(A) (B)
FIGURE 6‐2. A) CATENARY MOORING, B) TAUT MOORING. (ADAPTED FROM CHAKRABARTI[6‐5])

Naturally buoyant structures are usually subjected to substantial movement in the vertical and lateral
directions. Spars, however, have significantly less movement (around 3 m of vertical stroke) because a big part
of the hull is submerged (deep draft).

144
 Offshore Structures for Oil and Gas Production M. Stanko

Positively buoyed structures are moored vertically and keep a pre‐fix tension level. Whenever external loads
try to displace it, the mooring lines create a lateral tension that brings the structure back in place. The vertical
motion is therefore limited, but they are subjected to some displacement in the lateral direction.

6.1. SELECTION OF PROPER MARINE STRUCTURE

The selection of the marine structure to employ depends on multiple factors such as water depth, marine
loads, reservoir structure, soil conditions, future development plans, well artificial lift, among others. Some of
these factors will be described in more detail next.

6.1.1. WATER DEPTH

Figure 6‐3 shows the water depth range of the most common offshore structures for hydrocarbon production.
For shallow water depths (<450 m) the preferred and most economical option is usually a fixed structure such
as steel jacket or Gravity Based structure. For medium to deep waters, floating structures are preferred such
as TLPs, SPARs, FPSOs and Semi‐Subs. For ultra‐deep waters, FPSOs and Spars are more commonly used.

FIGURE 6‐3. WATER DEPTH RANGE OF THE MOST COMMON OFFSHORE STRUCTURES FOR HYDROCARBON PRODUCTION

6.1.2. LOCATION OF THE CHRISTMAS TREE

There are two main alternatives where to place the well trees: above (dry), or below (wet) the waterline. This
has a direct effect on the type of offshore structure to employ because only bottom‐supported structures,
TLPs and spars have a low enough motion range that is suitable for having dry trees. FPSOs and semi‐subs use
typically subsea wells (wet trees), the production is usually comingled with flowlines at the seabed and
transported with risers to the deck. Field developments might employ only dry trees, only wet trees of a
combination of both.

COMPLETION BITE: WELLHEAD ARCHITECTURE

The wellhead has the following main functions:
 Provides structural support (suspension point) for all casings and tubing strings. It transfers all loads
to the ground through the conductor.

145
 Offshore Structures for Oil and Gas Production M. Stanko

 It seals off each annulus at the top (at the bottom such seal is achieved by cementing). This is to avoid
leakages and to avoid that an outer casing, of a smaller pressure rating, will be exposed to full reservoir
pressure and therefore fail.
 Provide a connection point (interface) with the BOP and the Christmas tree.
 Provide annulus access and monitoring.
The procedure to deploy a wellhead during onshore drilling operations is described next. The focus is primarily
on the wellhead thus some details about the drilling process are omitted. The mechanical details of wellhead
components are simplified for clarity.
1. Dig the cellar, drill the conductor hole (typically 36 in), run the conductor (typically 30 in and length
40 m‐120 m) and cement it. Cut the conductor to the desired height (such that the production master
valve will be easy to operate at ground level).

FIGURE 6‐4. DEPLOYMENT OF THE CONDUCTOR

2. The BOP is placed, the surface casing hole is drilled (typically 24 in), the surface casing is run (typically
20 in) and cemented. The well is plugged and the BOP is removed. A baseplate is installed to transfer
all loads to the conductor and the casing head. The casing head is attached to the surface casing by
welding, threaded or with slips.

FIGURE 6‐5. RUN OF THE SURFACE CASING AND CASING HEAD

The casing‐casing head seal is positive‐pressure tested from below with the pressure port.

146
Offshore Structures for Oil and Gas Production M. Stanko

FIGURE 6‐6. DETAILS OF THE PRESSURE PORT ON THE CASING HEAD TO MAKE THE PRESSURE TEST

3. The BOP is placed, the intermediate casing hole is drilled (typically 17 ½ in), the intermediate casing is
run (typically 13 3/8 in) and cemented. The casing is hang on the casing head with the casing hanger
(set of slips, wedge, elastomer and no‐go shoulder).

FIGURE 6‐7. CASING HEAD WITH THE INTERMEDIATE CASING HANGED

The weight of the casing drives the slips down, presses the wedge that in turn squeezes the elastomer
and activates the seal. Lockdown screws (that pass through the flange, not shown in the figure) are
sometimes used to lock the upper part of the casing hanger and avoid unseating if the casing
experiences thermal expansion.

FIGURE 6‐8. DETAILS OF THE CASING HANGER (SLIPS AND SEALS)

The casing hanger can also be screwed, instead of using slips (also known as mandrel‐type hanger).
A negative pressure test is performed to ensure the casing hanger seal has been set properly.
4. The well is temporarily plugged, the BOP is removed and the casing spool for the intermediate casing
is installed (flanged). The casing hanger seal and the gasket are positive‐pressure tested from above
using the pressure port.

147
Offshore Structures for Oil and Gas Production M. Stanko

FIGURE 6‐9. INSTALLATION OF THE CASING SPOOL TO THE CASING HEAD

Steps 3 and 4 are repeated as many times as number of intermediate casings are planned for the well.

5. After all casings are hanged on the wellhead, the tubing head is bolted to the last casing spool.
The tubing is ran in hole and the tubing hanger is threaded to the last tubing joint. The tubing
is then hanged on the tubing head. The seal of the tubing hanger is activated with the
lockdown screws.

Depending on the application, the tubing hanger may have a port for hydraulic lines
(activation of SSSV, ICV), instrumentation line (pressure and temperature gauges), power lines
(ESP), etc.

FIGURE 6‐10. FINAL CONFIGURATION OF THE WELLHEAD

SAFETY STRATEGY FOR WELLS

 There must be two pressure barriers between the reservoir and the surface (in series)

148
 Offshore Structures for Oil and Gas Production M. Stanko

 No single event will compromise the two barriers at the same time (the barriers must be physically in
different places)
 The two barriers must be functionally independent.
 The barriers must be always tested from the direction of flow. (or using a negative pressure test).
 In case of barrier failure, the barrier must be reinstated as soon as possible.

The selection of dry or wet trees is further influenced by the spread of the reservoir, the future drilling or well
intervention plan and the water depth. For example, dry wells are preferred if it is feasible to produce a big
part of the reservoir with wells drilled from a single location. Also, if regular well intervention or recompletion
is expected during the life of the field. This is the case for example when wells are equipped with electric
submersible pumps that have a limited lifetime (around 2 years). Contrastingly, subsea wells usually require
intervention every 5 years.
Regarding water depth, dry tree systems have been used up to 1,700 m water depth.
If dry trees are used, the offshore structure has a well bay from where wells are drilled and completed and it
is equipped with a drilling package. It is also possible to have structures with drilling package and subsea wells
(e.g. Semi‐Sub Njord in the Norwegian Continental Shelf) where the wells are located exactly beneath the
structure. The size of the drilling package determines the drilling reach, a larger drilling package will allow to
drill longer wells but the structure must be bigger and therefore more expensive.
In steel jackets and GBSs wells are drilled and completed in a similar manner as onshore, wellheads and
Christmas trees are placed above the waterline. The well loads are supported by the soil and not the structure.
In TLPs and Spars, the wellhead is located on the seabed and there is a flex joint connector and a riser. The
riser ends further at the deck there is a secondary wellhead (with the tubing hanger) and the christmas tree.
There is a dynamic tensioner or buoyancy cans that support the tree and the production riser (Figure 6‐11a
and Figure 6‐11b).

(A) (B)
FIGURE 6‐11. TOP TENSION SYSTEMS FOR PRODUCTION RISERS IN FLOATING STRUCTURES (ADAPTED FROM CHAKRABARTI[6‐6])

149
 Offshore Structures for Oil and Gas Production M. Stanko

Other considerations to take into account are:

 Dry well structures have usually a limited number of well slots available due to space constraints and
load capacity. This makes the system less flexible for future expansion (infill drilling).
 Systems with subsea wells require special handling regarding flow assurance, to ensure the
uninterrupted transport of hydrocarbons in the flowlines from the seabed to the facilities.
 In systems with subsea wells, production can usually occur as soon as the facilities are commissioned.
New wells are tied‐in to the system as they become available.
 Fiscal metering requirements might affect the type of well to use. If the only test method allowed is
using a gravity vessel, then platform wells might be a better choice to avoid having several risers from
subsea wells.
6.1.3. OIL STORAGE
While gas is typically reinjected into the formation or sent through a transportation pipeline, oil is usually
transported from the field to the market using shuttle tankers. Sometimes it is desirable to store crude
temporarily on site to avoid stopping production in case of delays in the tankers’ trips due to external factors
(e.g. harsh weather conditions, remote locations). Table 6‐1 shows the storage capacities (qualitative) of the
most common offshore structures used for hydrocarbon production.

TABLE 6‐1. QUALITATIVE STORAGE CAPACITY OF COMMON OFFSHORE STRUCTURES

No or limited storage Steel Jackets, Semi‐subs, TLPs, Spars20
Medium ‐ Large
storage
FPSOs, GBS
(up to 2.5000.000
STB)

6.1.4. MARINE LOADS ON THE OFFSHORE STRUCTURE

Offshore structures are subjected to 3 main types of external loads: wind, waves and currents. These three
loads usually fluctuate with time and induce movements on the structure.

FIGURE 6‐12. WIND AND CURRENT LOADS ON AN OFFSHORE STRUCTURE

20
The Aasta Hansteen spar is the only spar up to date that has liquid (condensate) storing capacity of 150,000 STB.

150
 Offshore Structures for Oil and Gas Production M. Stanko

The types of movement exhibited by an offshore structure can be roughly classified depending if they are
floating or bottom‐supported. Floating structures display boat‐like motion with heave, yaw, sway, pitch, roll
and surge (Figure 6‐13a). Bottom fixed structures usually display deflections along its height similar to a long
bar recessed into the seabed (Figure 6‐13b).

(A) (B)
FIGURE 6‐13. EXAMPLES OF TYPICAL MOVEMENTS EXHIBITED BY OFFSHORE STRUCTURES

During the design and selection process of the offshore structure displacement and stresses must be
calculated based on the transient forces caused by wind, waves and currents. Some other considerations are:

 To determine the optimal location of flare of processing facilities to avoid fumes reaching the structure
considering preferential wind directions.
 To determine required deck elevation to avoid waves reaching facilities (usually based on a 100‐year
wave).
The computation of displacement and stresses with time in such structures is done typically using numerical
models (and often validated with scaled experimental prototypes). Forces are calculated from wave, wind and
current loads and applied on the structure. Due to the variability of these loads, there are usually three main
design approaches:

 Design wave: perform the analysis using the 100‐year significant wave height (HS,100) and a suitable
range of wave periods. If more accurate estimates are not available, the Norwegian standard NORSOK
N‐003 suggests to take HS,100 = 1.9 HS and vary the wave period between 6.5 ∙ 𝐻 , 𝑇
11 ∙ 𝐻 ,
 Short term design: perform the analysis for a 100‐year storm of specified duration (3‐6 h) with an
associated frequency spectrum. This is usually done to predict dynamic loads and stresses on critical
load‐bearing components.
 Long term design: This analysis takes into account the long‐term varying weather conditions. This is
important for fatigue design.
Resulting movement and stresses are time‐varying thus also must be analyzed statistically.
Every offshore structure has a “natural frequency” value that depends, simply put, on their weight, flexibility
and damping characteristics. If the structure is excited by external forces with a frequency that coincides with
the natural frequency, it will exhibit maximum amplitudes (a phenomenon called resonance). Correspondingly,
maximum amplitudes usually cause maximum loads and stresses on the structure thus most be avoided. This
is a phenomenon that occurs for all relevant movements the structure might have (described in Figure 6‐13).
A factor that is typically used in marine engineering is the Response Amplitude Operator (RAO). This value
gives the relationship between the amplitude of the response and the amplitude of the excitation for a range
of frequencies of the excitation force. As an example, consider the Heave RAO of a Sevan‐type FPSO presented

151
 Offshore Structures for Oil and Gas Production M. Stanko

in Figure 6‐14. The RAO gives the ratio of the amplitude in vessel heave by the wave amplitude for a range of
excitations periods. There is a very clear peak, around an 11 s period where the heave response is greatest.

FIGURE 6‐14. HEAVE RAO OF A SEVAN FPSO (TAKEN FROM SAAD ET AL. [6‐7])

Figure 6‐15 shows the natural periods of some offshore structures and the period range of some
environmental loads. Structures might be subjected to resonance if these two values coincide.

FIGURE 6‐15. ILLUSTRATIVE FIGURE INDICATING NATURAL PERIODS OF SOME OFFSHORE STRUCTURES AND EXCITATION PERIODS
OF SOME ENVIRONMENTAL LOADS

6.1. TREATMENT OF WIND, WAVES AND CURRENTS

Wind and current consist of flow velocity profiles along the vertical direction impacting on the structure (VW
and VC in Figure 6‐12). The magnitude of the velocity usually fluctuates in time (currents typically fluctuate

152
 Offshore Structures for Oil and Gas Production M. Stanko

with a period of hours and wind with a period between seconds and a minute). There are also some
preferential directions that exhibit stronger magnitudes than others (as shown in Figure 6‐16 for wind).

FIGURE 6‐16. WIND ROSE, (ADAPTED FROM HTTPS://SUSTAINABILITYWORKSHOP.AUTODESK.COM/BUILDINGS/WIND‐ROSE‐

DIAGRAMS)

During the design process of offshore structures, wind and currents are usually considered time invariant.
Wind is considered uniform while the variability of current with depth is usually accounted for. The value used
for design is the hundred‐year value (value that on average is met or exceeded only once in a hundred years
for a given location). An exception to this methodology is floating structures, where wind might have a more
relevant effect and its variability must be taken into account.
Waves are fundamentally variations of the sea level in space and time caused by wind. Figure 6‐17 shows a
wave time profile that displays a random behavior.

FIGURE 6‐17. TWO‐DIMENSIONAL RANDOM WAVE TIME PROFILE

Following Fourier’s theorem, this complex wave signal can be decomposed as the sum of “N” sine or cosine
functions each with an associated specific amplitude (ζai), frequency (ωi) and angle shift (εi). Please note that
the frequency is the inverse of the period.

153
 Offshore Structures for Oil and Gas Production M. Stanko

FIGURE 6‐18. CONTRIBUTION OF INDIVIDUAL REGULAR WAVES

Information about the components that make up a particular signal is commonly displayed in a wave energy
spectrum (Figure 6‐19a). The spectrum is the result of applying a Fast Fourier Transform (FFT) to the wave
signal and displays wave energy spectrum (Sζ) vs frequency (ω). For particular case of a regular wave made of
a single frequency component, the spectrum will display just a delta in the corresponding frequency. Common
spectrum formulas are Pierson‐Moskowitz (P‐M) and JONSWAP.

(A) (B)
FIGURE 6‐19. WAVE ENERGY SPECTRUM A) CONTINUOUS AND B) DISCRETIZED

The spectrum is often presented in a discretized manner (Figure 6‐19b) where the frequency axis is split in
segments (of width Δωi) and each segment has an associated frequency value (ωi) and wave energy value (Sζi).
The relationship between wave energy spectrum and wave elevation is shown in Eq. 6‐1.
EQ. 6‐1
𝜍 2 ∙ 𝑆 ∙ ∆𝜔

In short periods of time (typically 3 hours, called “sea state”) the spread in frequency is usually relatively low
(there is one dominant period called the “mean” period TZ) such that it is practically considered constant. The
wave elevation is assumed to follow a Gaussian type probability density function (Figure 6‐20a) and the wave
height (H, distance between consecutive peak and valley) a Rayleigh distribution (Figure 6‐20b).

154
 Offshore Structures for Oil and Gas Production M. Stanko

(A) (B)
FIGURE 6‐20. SHORT TERM PROBABILITY DENSITY FUNCTION OF WAVE ELEVATION (A) AND HEIGHT (B)

A parameter that is often used is the significant wave height Hs, defined as the average of the wave height in
the range between the maximum wave height (Hmax) and the wave height value for which 1/3 of all wave
heights are greater or equal than it (H1/3).
In design of offshore structures, it is desired to have accurate values of the expected wave characteristics
during their lifetime (long term statistics of ocean waves). Historic wind and wave data is typically gathered
with instrumentation located in weather and merchant ships, buoys, existing offshore structures, etc. To
establish statistical information of the sea state parameters, waves have to be measured for at least a couple
of years.
In such long periods of time, there will be a variation of the significant wave height and the mean period. The
data is often presented in a scatter diagram (an example shown in Figure 6‐21, gathered during a period of 15
years) that shows number of occurrences (Sea states) for a particular combination of significant wave height
and mean period. The mean period is also called spectral peak period (Tp). The particular case shown, the sea
displays a wave height between 1‐14 m and a period between 3‐20 s. The color red indicates combinations
that occurred only a few times, yellow medium and green combinations that were more frequent. Generally
speaking, in storms waves exhibit periods between 5‐25 s.

Spectral Peak period (Tp) [s]

Hs [m] 0‐3 3‐4 4‐5 5‐6 6‐7 7‐8 8‐9 9‐10 10‐11 11‐12 12‐13 13‐14 14‐15 15‐16 16‐17 17‐18 18‐19 19‐20 20‐21 21‐22 22‐23 23‐24 24‐25 Sum
0‐1 15 290 1367 2876 3716 3527 2734 1849 1138 656 362 192 101 52 26 13 7 3 2 1 0 0 0 18927
1‐2 1 81 1153 5308 12083 17323 18143 15262 10980 7053 4169 2316 1229 631 315 155 75 36 17 8 4 5 1 96348
2‐3 0 2 94 1050 4532 10304 15020 15953 13457 9752 5991 3403 1795 894 426 197 88 39 17 7 3 1 1 83026
3‐4 0 0 2 72 686 2782 6171 8847 9189 7493 5082 2991 1577 762 345 148 61 24 9 4 1 0 0 46246
4‐5 0 0 0 2 51 433 1645 3495 4807 4750 3638 2286 1229 584 251 100 37 13 5 1 0 0 0 23327
5‐6 0 0 0 0 2 39 294 1037 2069 2664 2440 1709 968 463 193 72 25 8 2 1 0 0 0 11986
6‐7 0 0 0 0 0 2 32 215 692 1264 1485 1228 767 382 159 57 18 5 1 0 0 0 0 6307
7‐8 0 0 0 0 0 0 2 27 157 447 730 762 555 302 130 46 14 4 1 0 0 0 0 3177
8‐9 0 0 0 0 0 0 0 2 23 112 276 392 355 223 104 38 11 3 1 0 0 0 0 1540
9‐10 0 0 0 0 0 0 0 0 2 19 77 160 192 148 79 31 9 2 0 0 0 0 0 719
10‐11 0 0 0 0 0 0 0 0 0 2 16 50 85 85 55 24 8 2 0 0 0 0 0 327
11‐12 0 0 0 0 0 0 0 0 0 0 2 12 29 40 33 18 7 2 0 0 0 0 0 143
12‐13 0 0 0 0 0 0 0 0 0 0 0 2 8 15 17 12 5 2 0 0 0 0 0 61
13‐14 0 0 0 0 0 0 0 0 0 0 0 0 2 5 7 6 4 1 0 0 0 0 0 25
14‐15 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 2 1 0 0 0 0 0 9
15‐16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 4
16‐17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
17‐18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sum 16 373 2616 9308 21070 34410 44041 46687 42514 34212 24268 15503 8892 4587 2143 921 372 146 55 22 8 6 2 292172

FIGURE 6‐21. SCATTER DIAGRAM OF LONG TERM WAVE STATISTICS

A number that is typically reported and used during the design process of offshore structures is the significant
wave height that might be reached or exceeded during a period of 100 years (HS,100). As data hasn’t been
collected for such long periods of time, an extrapolation of the wave data collected in the scatter diagram is
performed. The extrapolation is done using a Semi‐logarithmic distribution that relates the significant wave
height versus the chance of exceedance.

155
 Offshore Structures for Oil and Gas Production M. Stanko

1 EQ. 6‐2
log 𝑃 𝐻 ∙𝐻
𝑎
The 100‐year period is constituted by 292,000 “sea states” of 3 h duration (where the significant wave height
and the period can be considered constant). The 100‐year wave occurs is reached or exceeded only once, thus
its probability is 1/292,000 i.e. 3.40 x 10‐6.
The data shown in Figure 6‐21 has been gathered during a period of 15 years (probability of occurrence of a
15 years wave is 2.3 x 10‐5). If one particular spectral period range is chosen (e.g. 18‐19) then the probability
density of the wave height and the cumulative distribution can be computed. The significant wave height that
will likely occur once in 15 years is then can be read from the cumulative distribution (16.5 m).

(A) (B)
FIGURE 6‐22. PDF AND CD OF SIGNIFICANT WAVE HEIGHT FOR SPECTRAL PERIOD RANGE 18‐19 S

Then, it is possible to compute a from Eq. 6‐2 (for the particular case a = ‐3.5531). The significant wave height
of 100 years is then computed with Eq. 6‐2 and a probability of 3.40 x 10‐6. The 100‐year wave obtained is 19.4
m for the period range 18‐19 s for the particular case. This value is often used to determine the required
distance between the deck and the sea level.

156
Offshore Structures for Oil and Gas Production M. Stanko

REFERENCES
[6‐1] Petrowiki. Offshore and Subsea Facilities. SPE publications. Retrieved from
http://petrowiki.org/Offshore_and_subsea_facilities, on Jan 9th, 2017.
[6‐2] Petrowiki. Compliant and Floating systems. SPE publications. Retrieved from
http://petrowiki.org/Compliant_and_floating_systems, on Jan 9th, 2017.
[6‐3] Offshore magazine. 2015 Deepwater Solutions and Records for Concept Selection. Poster. Retrieved
from http://www.offshore‐mag.com/content/dam/offshore/print‐articles/volume‐75/05/0515‐
DeepwaterPoster040815ADS.pdf, on Jan 9th, 2017.
[6‐4] Stell, J. Wet tree vs. dry tree criteria. OE Offshore engineer. Retrieved from
http://www.oedigital.com/component/k2/item/9601‐wet‐tree‐vs‐dry‐tree‐criteria, on Jan 9th,
2017.
[6‐5] Chakrabarti, S. (2005). Handbook of Offshore Engineering. Volume I. Elsevier.
[6‐6] Chakrabarti, S. (2005). Handbook of Offshore Engineering. Volume II. Elsevier.
[6‐7] Saad, A. C., Vilain, L., Loureiro, R., Brandao, R.M., Filho, R.Z. Motion behavior of the Mono‐Column
FPSO Sevan Piranema in Brazilian Waters. OTC paper 20139. Offshore Technology Conference,
May, 2009.

157
 Flow Assurance Management in Production Systems M. Stanko

7. FLOW ASSURANCE MANAGEMENT IN PRODUCTION SYSTEMS

Flow assurance consists in ensuring uninterrupted flow of hydrocarbon streams from the reservoir to the point
of sale according to production plan. Flow assurance is particularly relevant for deep subsea systems with
relatively long transportation distances (5‐150 km) and low surrounding temperatures. If there is a problem
intervention and remediation, these activities usually must be done remotely and are time consuming and
expensive.
Flow assurance focuses on three main aspects:
1. Avoid flow restrictions (excessive pressure drop, blockage or intermittent production).
2. Safeguard the structural integrity of parts of the production system from damages caused by internal
flow.
3. Maintain the functionality and operability of components in the production system.
There are multiple issues that are typically addressed in flow assurance:

 Formation and deposition of wax.

 Formation of hydrates.
 Formation and accumulation of scale
 Flow induced vibrations (FIV)
 Asphaltene formation and deposition
 Slugging
 Erosion
 Emulsion
 Corrosion
 Pressure surges during shutdown and startup.
 Naphtenates
 Foaming
 Liquid loading
Figure 7‐1 shows where these issues usually occur in the production system.

RESERVOIR WELL PIPELINES PROCESS

ASPHALTENES

SCALE

HYDRATES

WAX

EMULSIONS

FIV

CORROSION

SLUGGING

EROSION

LIQUID LOADING

NAPHTENATES

FOAM

MICROORGANISM GROWTH

FIGURE 7‐1. FLOW ASSURANCE PROBLEMS AND THEIR TYPICAL LOCATION IN THE PRODUCTION SYSTEM

158
 Flow Assurance Management in Production Systems M. Stanko

An overview of some of these flow assurance issues is provided next.

7.1. HYDRATES

Hydrates are solid substances where water molecules (in liquid phase) form a cage‐like structure that hosts
small (< 9 Å diameter) molecules (Figure 7‐2). The small molecules are usually methane, ethane, propane,
butane, carbon dioxide, nitrogen. The cage‐type structure is formed due to hydrogen bonding of water
molecules (the water molecule tends to spacially create two positives and a negative pole).

(A) (B)
FIGURE 7‐2. A) APPEARANCE OF A HYDRATE PLUG (PHOTO TAKEN FROM SCHROEDER ET AL[7‐1] ), B) MOLECULAR STRUCTURE OF A
METHANE HYDRATE

Hydrates contains a much higher proportion of water than the hydrocarbon component. For example, a
methane hydrate (called methane clathrate) with molecular formula 4CH4∙23H2O (MW = 478 kmol/kg) has a
molar proportion of 85% (23/27) water and 15% (4/27) methane.
However, this does not necessarily indicate that they contain small amounts of gas. For example, one cubic
meter of methane clathrate (of an approximate density of 900 kg/m3) contains 1.88 (900/478) kmol of hydrate,
of which there are 7.53 (1.88 x 4) kmol of methane. 7.53 kmol of methane at standard conditions correspond
to 178.4 Sm3! (VSC = nmoles·R·TSC/pSC). For a cubic meter to contain the same amount of gaseous methane at
standard temperature, it would have to be compressed at 180.4 bara (p= 7.53 kmol·R·TSC/1 m3).
Hydrates form only if ALL following ingredients are present:
 Free water (in liquid phase)
 Small hydrocarbon molecules
 Particular range of pressure and temperature.
An example of the hydrate formation region is shown in Figure 7‐3. The actual line depends mainly on the fluid
composition, but, as a rule of thumb, it happens at high pressure and low temperatures. For example, at a
pressure of 12 bar, the hydrate formation temperature is 4 °C.

159
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐3. HYDRATE FORMATION REGION

The hydrate formation line can be predicted by empirical expressions (that are a function of the specific gravity
of the gas) or using equilibrium calculations with an Equation of State. Hydrate equilibrium calculations
resemble to Vapor Liquid equilibria by finding p and temperature conditions that make equal the chemical
energy of the component in the hydrate phase and liquid and gas phases.

7.1.1. CONSEQUENCES
If the pressure and temperature of the fluid flowing along the production system falls inside the hydrate
formation region, hydrates will start to form. Hydrates usually form at the liquid‐gas interphase where free
water and small hydrocarbon molecules are in contact. The mixing and turbulence of the flow further
increases the contact between the two thus causing the formation of more hydrates. Hydrates then start to
agglomerate until they eventually plug the pipe (Figure 7‐4).

1 * 2

FIGURE 7‐4. EVOLUTION OF P AND T OF THE FLUID WHEN FLOWING ALONG THE PRODUCTION SYSTEM

Hydrates can also form when the production is stopped and the stagnant fluid begins to cool by transferring
heat with the environment.

7.1.2. MANAGEMENT
The traditional strategy to manage hydrates is to avoid their formation. There are two main techniques
commonly used to prevent the formation of hydrates:

160
 Flow Assurance Management in Production Systems M. Stanko

 Keep the fluid conditions out of the hydrate formation region. This is done mainly by reducing the rate
of temperature drop of the fluid (reducing the lateral spread of the blue line in Figure 7‐4). This is
achieved in practice by two methods: better insulation or electrical heating of the pipe.
Please note that insulation works effectively for a flowing system, but when production is stopped, usually
some other control method must be used as the fluid will eventually cool down during a long period.
Electrical heating is usually not cost effective for long transportation distances.

 Reduce the hydrate formation region. The equilibrium pressure and temperature of hydrate formation
can be affected by adding liquid inhibitors (typically Mono‐ethylene‐glycol MEG, Tri‐ethylene‐glycol
TEG or methanol MEOH) to the water phase. Inhibitors interfere with the formation of hydrogen
bonds by keeping water molecules apart. As a consequence, the hydrate formation line will be shifted
to the left (as shown in Figure 7‐5).

FIGURE 7‐5. EFFECT OF INHIBITOR INJECTION ON THE HYDRATE LINE

Typical concentrations of inhibitors used are 30‐60 in weight %. For example, the Snøhvit field has a Water
Gas ratio of 6.00 x 10‐6 Sm3/Sm3. The plateau production of the field is 20 MSm3/d, thus it produces around
120 Sm3/d of water, or, equivalently, 120,000 kg/d of water. If we assume that the inhibitor concentration
used is 50 in weight %, then this gives 120,000 kg/d of MEG that must be continuously injected on the field.
This represents a daily cost of 60,000 – 180,000 USD (assuming a MEG cost between 0.5 – 1.5 USD/kg).
Therefore, MEG is usually reclaimed in the processing facilities.
The inhibitor must be present in the water phase for it to be effective, thus evaporation to the gas phase has
to be taken into account when estimating the required amounts of inhibitor.
Inhibitors are also injected when preparing to shut down production, to make sure hydrates will not form due
to the cooling of the fluid.
Figure 7‐6 shows a flow schematic of a subsea production system highlighting the hydrate inhibitor injection
system (in green color). The production system has 2 satellite wells, a manifold template, and two pipelines
that transport reservoir fluids topside. The hydrate inhibitor is transported with an umbilical21 from topside
until the subsea distribution unit. In the subsea distribution unit, the hydrate inhibitor line in the umbilical is
connected to a distribution manifold that is further connected to the wells with a separate flowline (if the
distance is short) or with an umbilical.

21
The umbilical is a flexible pipe‐like structure that encloses other pipes that transport chemicals, hydraulic fluid to
actuate valves, electrical cables, fiber optic lines, etc.

161
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐6. FLOW SCHEMATIC OF A SUBSEA PRODUCTION SYSTEM WITH HYDRATE INHIBITOR INJECTION SYSTEM

Figure 7‐7 shows in more detail the pipe and cable splitting that occurs in the subsea distribution unit, the
distribution manifold for the hydrate inhibitor and the isolation valves for each well (ROV operated). Figure
7‐8 shows in more detail how is the hydrate inhibitor injection system integrated with the well tree.

FIGURE 7‐7. DETAILS OF A SUBSEA DISTRIBUTION UNIT.

162
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐8. HYDRATE AND SCALE INHIBITOR INJECTION SYSTEM IN THE X‐MAS TREE

In the last years, experts have proposed to use a less conservative hydrate control strategy where we allow
hydrates to form but impede their agglomeration and carry the slurry together with the production fluids. This
can be performed by injecting special types of chemicals, or by using cold flow. However, up to date there are
limited field cases where this type of management is performed.

7.2. SLUGGING
Slugging consists on intermittent flow of gas and liquid in the production system (Figure 7‐9).

FIGURE 7‐9. SLUG IN A PIPE SECTION

There are two main types of slugging:

 Hydrodynamic slugging: It occurs spontaneously at a particular combination of flow velocities of liquid

and gas and it depends strongly on the fluid properties and pipe inclination. As an example, Figure
7‐10 shows the flow pattern map for a horizontal pipe and certain fluid properties. There is a particular
combination of operational velocities where the flow will arrange itself in a slug flow configuration.

163
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐10. FLOW PATTERN MAP FOR A HORIZONTAL PIPE

 Terrain slugging: Terrain slugging is mainly due to cyclic accumulation of liquid in the production
system (especially in lower points). This happens in undulating well trajectories, transportation
flowlines with varying topology of the seabed and in risers.
An example of slugging in a s‐shaped production riser is shown in Figure 7‐11. Liquid accumulates in the lowest
pipe section and blocks the flow of gas (a). The liquid level starts increasing and the gas pressure in the
horizontal line also increases (b). Eventually, the liquid floods the second floor of the riser (c). Gas pressure
increases until it is sufficient to flush out almost all the liquid in the riser (d).

(A) (B) (C) (D)

FIGURE 7‐11. STAGES OF SEVERE SLUGGING IN AN S‐SHAPED RISER

7.2.1. CONSEQUENCES
The main consequence of slugging is that production rates and pressures will fluctuate in time which is often
detrimental to the proper operation of the downstream processing facilities. In gravity separators for example,
a sudden inlet of liquid might increase significantly the liquid level, causing liquid carryover, activating the
warnings for high liquid level and even triggering a shutdown alarm. The distance between the normal liquid
level and the high alarm level should be big enough to accommodate the volume of the biggest liquid slug
expected.
Slugging also causes vibration in flowlines, manifolds, risers which can develop in structural damages due to
elevated stress levels and fatigue.

7.2.2. MANAGEMENT
Slugging can be, to some extent, predicted during the design phase of the field using commercial multiphase
flow simulators such as LedaFlow, Olga and FlowManager. If it is detected and it has high severity (long slug
lengths, frequencies that coincide with the natural frequency of the structure, relevant pressure fluctuations),
potential solutions are to change the routing of the flowline, refill or dig some sections of the seabed that can
cause liquid accumulation or changing the pipe diameter. Smaller pipe diameters increase the gas velocity,

164
 Flow Assurance Management in Production Systems M. Stanko

increasing the drag of the gas on the liquid thus reducing the liquid deposition. However, too‐small pipe
diameters also cause higher pressure drops that reduce overall production rates.
If slugging is occurring in an existing production system, some approaches that have been used successfully in
the past are to apply gas lift in the riser base or to use the topside choke to change dynamically the
backpressure on the line and “control” the slug.

7.3. SCALING
Scaling is the precipitation of minerals compounds (constituted by Na, K, Mg, Ca, Ba, Sr, Fe, Cl) from the
produced water and their deposition on pipe walls. Scale occurs when the solubility of the minerals in the
water decreases due to changes in pressure and temperature, due to mixing of waters of different sources,
injection of CO2. Minerals usually deposit on surface areas that are rough or have irregularities (e.g. valve
components).

FIGURE 7‐12. SCALE ACCUMULATION IN CHOKE (IMAGE TAKEN FROM SANDENGEN[7‐2] )

There are two main types of scales that usually occur in production systems:
 Carbonate scales. These scales are formed when CO2 dissolved in the water disassociates in carbonate
ions CO3‐2 and join with some of the aforementioned minerals (typically calcite CaCO3, Iron
carbonate FeCO3). Their precipitation is mainly due to reduction in pressure (due to flow in
restrictions, valves, chokes) or increases in temperature. This type of scale can be removed with acid.
 Sulphate scales: These scales are formed by the sulphate ion SO4‐2 that is present in seawater (Barite
BaSO4, Gypsum CaSO4∙2H2O, Anhydrite CaSO4, Celestite SrSO4). It precipitates out of solution when
waters from different sources are mixed (e.g. seawater used for injection and production water from
the aquifer or formation). The pressure has little influence in the precipitation, but the increase in
temperature can reduce further the solubility. This type of scale must be removed mechanically.
7.3.1. CONSEQUENCES
Scaling causes gradual blockage of the flow path and loss of functionality in production equipment (Subsurface
safety valves, chokes).

7.3.2. MANAGEMENT
Studies are usually performed on the produced water to determine if it will be prone to form scale at the
pressure and temperature conditions encountered in the production system. Moreover, special attention
must be payed to situations where there is mixing of water from different sources, CO2 injection.
Scaling is usually avoided by using chemicals (scale inhibitors) that attach themselves to the scale ions and
impede growth. Coating can help to prevent deposition on the surfaces but when damaged (e.g. due to
erosion) their effectivity is reduced dramatically.

165
 Flow Assurance Management in Production Systems M. Stanko

If scale forms in a component of the production system, the removal technique depends on the type of scale.
Carbonates can be removed by acid injection and sulphates can only be removed mechanically.

7.4. EROSION
Erosion is the gradual damage and loss of material from the wall of components of the production system
(valves, pipes, bends, etc. Figure 7‐13) due to the repeated impingement of solid particles (sand) or droplets
at high velocity.

FIGURE 7‐13. EROSION DAMAGE IN A CAGE‐TYPE CHOKE [SOURCE UNKNOWN)

7.4.1. CONSEQUENCES
Structural damage, vibration, leaks and corrosion (due to the removal of the protective coating).

7.4.2. MANAGEMENT
Erosion is usually accounted for in the field design phase. The design process sizes the equipment such that
the velocities are below certain limit value that gives an acceptable erosional rate. These calculations usually
consider the velocity of impingement, the angle of impingement, the concentration of solid particles and the
wear resistance of the material.
There are some standards that give guidelines how to estimate erosive wear for common pipe components
(e.g. DNV Recommended Practice RP O501). However, complex geometries usually require in‐depth studies
(e.g. using computational fluid dynamics, CFD) to estimate erosion‐prone areas, fluid velocities, angle of
impingement, etc. An example is shown in Figure 7‐14.

FIGURE 7‐14. CFD SIMULATION OF EROSION IN A PRODUCTION HEADER

If erosion is detected in an existing production system then, when possible, components might be reevaluated
and replaced with geometries that are less susceptible to erosion. Alternatively, if corrosion is due to excessive
sand production from the reservoir, the only alternative is then reducing the well rate to limit sand production.

166
 Flow Assurance Management in Production Systems M. Stanko

7.5. CORROSION
Corrosion is an electrochemical reaction where steel is converted to rust and occurs when metal is in contact
with water. Two locations are established in the metal, a cathode and an anode. In the anode, iron loses
electrons and becomes a positively charged ion. This ion further reacts with water and oxygen in the
surrounding media to form rust. The cathode receives the electrons of the anode and generates by‐products
(such as hydrogen H2) with other ions.

(A) (B) (C)

FIGURE 7‐15. A) ILLUSTRATION OR A CORROSION REACTION B) CORROSION ON A CASING INSIDE SURFACE C) CORROSION ON
TUBING

Corrosion can occur virtually anywhere in the production system where water is in contact with metal (casing,
tubing, flowlines, pipelines, tanks, pumps, etc.). In transportation pipes, corrosion usually occurs at the pipe
bottom where water is transported, in low pipe sections where water accumulates or at the top of the pipe
due to splashing and condensation of water droplets (also known as TLC, Top of line corrosion).

FIGURE 7‐16. WET GAS FLOW IN A HORIZONTAL FLOWLINE DEPICTING TOP OF LINE CONDENSATION

Oxygen and acid gases such as CO2 and H2S contribute to corrosion.

7.5.1. CONSEQUENCES
Corrosion on an unprotected pipe can cause losses of 1‐20 mm of pipe thickness per year, leading ultimately
to structural damage and leakages. Rust particles can also travel downstream and cause problems such as
plugging other components.

7.5.2. MANAGEMENT
The measures to mitigate corrosion can be divided into two main principles:

 Eliminate the contact between water from steel. This can be done by applying a protective layer on
the steel surface, for example with coating (which might be eventually damaged due to sand erosion),
creating a layer of protective oxide on the steel (Figure 7‐17a) or by using inhibitors (Figure 7‐17b).

167
 Flow Assurance Management in Production Systems M. Stanko

(A) (B)
FIGURE 7‐17. PROTECTIVE LAYER OF FECO3 FORMED ON THE METAL SURFACE B) INHIBITORS ATTACHED TO THE METAL SURFACE

 Use steel materials with higher resistance to corrosion. For example, alloy steels. This is usually feasible
for wells, but it becomes too expensive for flowlines and pipelines.

7.6. WAX DEPOSITION

Wax deposition occurs when long alkane chains (C18+) precipitate out of solution from the oil, agglomerate
and deposit on the pipe walls.
In a waxy crude, when temperature is reduced down to a certain value (for North sea crudes this happens
around 30‐40 C), some wax crystals will start to precipitate and become visible. The temperature when this
occurs is called cloud point or WAT (wax appereance temperature).

(A) (B)
FIGURE 7‐18. A) WAX CRYSTALS VISIBLE IN A CRUDE AT WAT, B) WATS AT DIFFERENT PRESSURES IN THE PHASE DIAGRAM

The WAT depends on oil composition, type and molar amounts of alkanes, pressure, cooling rate. Wax crystals
typically attach to nucleating agents present in the oil (asphaltenes22, fine sand, clay, water, salt), form wax
“clusters” and grow.
If the temperature is reduced further down to the pour point, the oil becomes solid‐like and stops flowing by
gravity.

22
Asphaltenes are coal‐like solids that also have the tendency to precipitate out of the crude. They are high molecular
weight compounds containing poly‐aromatic carbon rings with nitrogen, sulphur, oxygen and heavy metals such as
vanadium and nickel.

168
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐19. CRUDE OIL NOT FLOWING ONCE THE POUR POINT IS REACHED

Wax deposition occurs when ALL the following ingredients are present:
 Wax‐prone components in the oil composition (long alkane chains).
 Temperature below WAT.
 Pipe wall colder than the fluid such that there is a temperature profile in the fluid reducing towards
the pipe wall (temperature gradient).
 Irregularities on the wall where wax clusters attach.
Wax deposits age with time and become more rigid (thus more difficult to remove).

7.6.1. CONSEQUENCES
In flowlines and pipelines:
 Increases pressure drop due to the increase in pipe roughness.
 Reduction of cross section area.
 Pipe blockage.

(A) (B)
FIGURE 7‐20. A) WAX PLUG RETRIEVED TOPSIDE (IMAGE TAKEN FROM LABES‐CARRIER ET AL[7‐3] ), B) EVOLUTION OF THE WAX
THICKNESS IN A PIPELINE WITH TIME

 The presence of wax crystals in the fluids changes its rheology (e.g. making it non‐Newtonian or with
a higher effective viscosity).
 During shut‐downs, the temperature of the fluid can reach the pour point of the crude, causing it not
to flow (gelling).

169
 Flow Assurance Management in Production Systems M. Stanko

7.6.2. MANAGEMENT
The first step in developing a wax management strategy is to test the crude oil in the laboratory and measure
and quantify all of its properties relevant for deposition.
A common management method for wax is to perform frequent pigging. Pigging consists in sending a device
(pig) inside the pipe that scraps the wax deposits and pushes them forward. Pigs are usually sent and received
from the processing facilities thus two pipelines must be installed. There are also subsea pig launchers, but
this is economic only for systems with very low pigging frequency.
Figure 7‐21 shows the flow schematics of a subsea production system with two satellite wells producing to a
subsea manifold. There are two pipelines from the manifold to topside and there is a crossover valve on the
subsea manifold (normally closed) that allows to communicate both. When performing pigging operations,
the crossover valve is open and the pig is send through one pipeline with a pig launcher topside and received
through the other end, on the pig receiver.

FIGURE 7‐21. FLOW SCHEMATIC OF A SUBSEA PRODUCTION SYSTEM WITH FACILITIES FOR PIGGING AND INDIVIDUAL WELL
TESTING

Pigging frequency is usually estimated by performing numerical simulations to compute the profile of
deposited wax along the flowline with time. With this, the total amount of wax deposited in the system at any
given time is estimated. There is a maximum length and weight of wax that can be pushed through the pipe,
given by the maximum allowable pressure that the pipe can tolerate. The required pigging frequency is given
by the time at which that wax amount is reached.

170
 Flow Assurance Management in Production Systems M. Stanko

Other techniques used are keeping the fluid outside of the wax formation region. This is done by thermal
insulation or electrical heating. However, for long flowlines, electrical heating is usually very expensive and
insulation alone is not enough to keep temperature sufficiently high. Thus, in most cases insulation or electrical
heating are often used to reduce wax deposition rates together with pigging.
Chemical inhibitors that are also often injected. Chemical inhibitors work by reducing the cloud point of the
crude or by preventing further agglomeration of wax crystals. As with insulation, in many systems this does
not eliminate completely the problem but it helps slowing down the deposition rate. Chemical inhibitors are
typically expensive.
If the seabed temperature is below or equal the pour point of the oil, then it is necessary to inject chemical
inhibitors before shutting down the system to avoid gelling.
In recent years pipe coating has been proposed as a technique to avoid wax attaching to pipe walls. However,
it is not yet field tested.
In systems with wax‐prone oils the pressure drop between end points of flowlines should be closely monitored.
Any unexplained increase might indicate wax deposition and must be immediately addressed.

7.7. OIL‐WATER EMULSIONS

Oil‐water emulsions are fine and stable dispersions of oil droplets in water or water droplets in oil (Figure
7‐22). The formation of emulsions depends on a variety of factors such as the dynamics of multiphase flow,
the properties of oil and water such as viscosity and interfacial tension, the shear stress (mixing) experienced
by the mixture, chemical compounds present in the oil‐water interface. In production systems, the mixing is
typically generated when commingling production from different sources, due to the violent expansion across
the choke, flow through multiphase pumps, etc.

(A) (B)
FIGURE 7‐22. A) OIL (RED) AND WATER (WHITE) ORIGINALLY SEPARATED, B) OIL AND WATER EMULSION AFTER VIGOROUS
STIRRING IN A BLENDER. PHOTOS TAKEN BY HONG[7‐4]

7.7.1. CONSEQUENCES
In pipe flow, emulsions often exhibit the behavior presented in Figure 7‐23. For a fixed volumetric rate of the
mixture (qo + qw), if one measures the pressure drop along a pipe segment for several water volume fractions,
it will increase with water volume fraction until a maximum is reached and then it will decline abruptly. The
water volume fraction that has the highest‐pressure gradient is called the inversion point. The increase in
pressure drop is usually significant (more 2.5 times the one for pure oil in the figure).
When increasing the water fraction, at the inversion point the dispersion changes from a water in oil dispersion
to an oil in water dispersion.

171
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐23. MEASURED PRESSURE DROP IN A HORIZONTAL PIPE KEEPING THE TOTAL FLOW RATE CONSTANT AND CHANGING
WATER VOLUME FRACTION, 𝒒𝒘/ 𝒒𝒘 𝒒𝒐

Figure 7‐24 shows an oil‐water flow pattern map depicting mixture velocity (total liquid rate divided by pipe
cross section area) in the “x” axis, water cut in the “y” axis and the flow pattern regions in colors. The transition
shown in Figure 7‐23, from water volume fraction of zero to one at constant flow rate (mixture velocity) is
plotted as a vertical line on the figure at mixture velocity approximately equal to 0.5 m/s (arbitrary value).
Along the line the flow pattern changes from a dispersion of water in oil (Dw/o) to a dispersion of oil in water
(Do/w) and the inversion point occurs at a water volume fraction of around 0.5.

FIGURE 7‐24. OIL‐WATER FLOW PATTERN MAP OF WATER VOLUME FRACTION VERSUS MIXTURE VELOCITY FOR AN UPWARD PIPE
INCLINATION OF 45°. FIGURE ADAPTED FROM RIVERA[7‐5] [7‐1] .

Using a homogeneous model (single fluid with average properties) one can back‐calculate the effective
mixture or “emulsion” viscosity that the mixture should have to provide the pressure drop measured (Figure
7‐25). For the particular case, the emulsion viscosity at the inversion point (570 cP) is 7.1 times the viscosity
of the oil (80 cP).

172
 Flow Assurance Management in Production Systems M. Stanko

FIGURE 7‐25. MIXTURE VISCOSITY BEHAVIOR VERSUS WATER VOLUME FRACTION EXHIBITED BY THE OIL WATER MIXTURE

There are many expressions used to represent the behavior shown in Figure 7‐25 that are later used in
emulsion pressure drop models. Most of them require data measured in the lab to tune their coefficients. As
an example, the Richardson model is shown below.
∙ EQ. 7‐1
For oil continuous 𝜇 𝜇 ∙𝑒

𝜇 𝜇 ∙𝑒 ∙ EQ. 7‐2
For water continuous

Emulsions can cause excessive pressure drops in pipe segments and components, which can reduce
dramatically production rates, pumping capacity of electric submersible pumps, etc. Moreover, stable
emulsions are difficult to separate in processing facilities thus creating bottlenecks and fluid disposal problems.

7.7.2. MANAGEMENT
During the field design phase, the capacity oil and water system to form emulsions can be somewhat studied
with laboratory tests (shaking bottle tests). However, these results have sometimes limited applicability partly
because the shear magnitudes (mixing) applied in the laboratory conditions are very different from the mixing
experienced in the field.
When there is mixing of streams with different water cut, the inversion point must be avoided.
Often, chemical substances such as demulsifiers and light oils (diluent) are injected into the stream to reduce
the stability of the emulsion. Light oils reduce the viscosity of the formation oil, thus helping separation.
Demulsifiers are chemicals that attach themselves to the interface between oil and water promoting
separation.

7.1. SUMMARY TABLE

A summary table is provided describing briefly causes, potential consequences, prevention and solution
measures and tools available for analysis for some flow assurance issues mentioned above.

173
 Flow Assurance Management in Production Systems M. Stanko

TABLE 7‐1. SUMMARY TABLE OF FLOW ASSURANCE ISSUES: CAUSES, POTENTIAL CONSEQUENCES, PREVENTION AND SOLUTION
MEASURES AND TOOLS AVAILABLE FOR ANALYSIS

7.1. ABOUT CHEMICAL INJECTION

Many of the preventive and corrective measures against flow assurance issues involve the continuous or
occasional injection of chemicals and substances to inhibit the precipitation or dissolve solids. Such chemicals
often cannot be recovered and reused, but rather follow the produced oil, gas or water. Some of the chemicals
employed are often damaging to the environment and therefore their usage must be strictly controlled,
especially if they might end up in the environment (e.g. follow the disposed water).
In offshore installations, chemicals are classified in categories following applicable regulation (for example the
OSPAR, convention for the protection of the marine environment of the North‐East Atlantic). For example,
color codes are used to classify substances:
 Green: substances which pose little or no risk to the environment.
 Yellow: substances that are not classified as green, red or black. Typically includes substances with
low toxicity and that can be significantly degraded after 28 days.
 Red: substances that can accumulate in the environment and that have slow degradation times.
Requires special permission to use and discharge to environment.

174
 Flow Assurance Management in Production Systems M. Stanko

 Black: substances that do not degrade, that are poisonous and that accumulate in the environment.
Permission to use and discharge is only given due to safety or critical technical reasons.
Some examples of red chemicals are emulsion inhibitors, wax inhibitors, anti‐foamers. However, these
substances are injected or accumulate main in the oil, and therefore will not end in the environment.
Some examples of yellow chemicals are scale inhibitors, biocides. These substances are soluble in water and
will follow injection water or disposed produced water.
MEG is often classified a green‐type substance. However, it is more economic to recover it from the production
water for reuse.

175
Flow Assurance Management in Production Systems M. Stanko

REFERENCES
[7‐1] Schroeder, Jr.; Chitwood, J. Krasin, T.; Lee, B.; Krohn, D.; Huizinga, M.; Paramonoff, A.;Schroeder, C.;
Gay, T.; Cercone, D.; Pappas, J. Development and Qualification of a Subsea 3000 Barrel Pressure
Compensated Chemical Storage and Injection System. OTC paper 26966. Offshore Technology
Conference, May, 2016.
[7‐2] Sandengen, K. Hydrates and Glycols MEG (Mono Ethylene Glycol) Injection and Processing.
Presentation at NTNU. September, 9. 2010.
[7‐3] Labes‐Carrier, C.; Rønningsen, H. P.; Kolnes, J.; Leporcher, E. Wax Deposition in North Sea Gas
Condensate and Oil Systems: Comparison Between Operational Experience and Model Prediction.
SPE paper 77573. SPE Annual Technical Conference and Exhibition. San Antonio, Texas. 2002.
[7‐4] Hong, C. Study on Ultrasonic Influence on Oil‐Water Emulsion Separation. Project Report. Norwegian
University of Science and Technology. 2017.
[7‐5] Rivera, R. Water Separation from Wellstream in Inclined Separation Tube with Distributed Tapping.
PhD Thesis. Norwegian University of Science and Technology. 2011.

176
 Heat Transfer for Flow in Conduits M. Stanko

8. HEAT TRANSFER FOR FLOW IN CONDUITS

The equation for conservation of energy for a section of a conduit is
EQ. 8‐1
𝑄 𝑊 𝑚∙ 𝑒 𝑒

The specific energy that the stream has is usually split in internal energy (u), potential energy (z∙g) and kinetic
energy (V2/2). In this equation, the sign convention for 𝑄 and 𝑊 is when entering the pipe are positive, and
when leaving the pipe are negative.
A conduit doesn’t exchange work with the surroundings, but the fluid must perform work to enter and leave
the system. This specific work is: 𝑝 ∙ 𝑣 𝑝 ∙𝑣 (Here 𝑣 is specific volume).
By combining the inlet and outlet specific internal energy “u” with the specific work to enter and leave the
system to obtain specific enthalpy, the energy conservation equation is written as:

𝑉 𝑉 EQ. 8‐2
𝑄 𝑚∙ ℎ 𝑧 ∙𝑔 ℎ 𝑧 ∙𝑔
2 2
Or, alternatively

𝑉 𝑉 EQ. 8‐3
𝑄 𝑚 ∙ ∆ℎ ∆𝑧 ∙ 𝑔
2 2
Here ∆ represents outlet minus inlet.
In differential form (for an infinitesimally small length of pipe) the equation can be expressed as follows:

𝑑𝑄 𝑑ℎ 𝑑𝑧 𝑑𝑣 EQ. 8‐4
𝑚∙ 𝑔∙ 𝑣∙
𝑑𝐿 𝑑𝐿 𝑑𝐿 𝑑𝐿
Heat leaving the system is negative (the temperature of the outlet fluid is lower than the temperature at the
inlet and the term ∆ℎ is usually negative). Heat entering the system is positive.

8.1. ORDER OF MAGNITUDE ANALYSIS ON THE SPECIFIC ENERGY TERMS OF A STREAM

The energy conservation shown above is usually simplified by neglecting the changes in specific kinetic and
potential energy. This is because changes in specific enthalpy are usually much larger than changes in specific
kinetic and potential energy. To analyze this, we will compare the following three terms:

𝑉 𝑉 EQ. 8‐5
∆ℎ, ∆𝑒 ∆𝑧 ∙ 𝑔, ∆𝑒
2
Where ∆𝑒 and ∆𝑒 are changes in specific potential and kinetic energy respectively.
To perform this comparison, we will use as a reference a vertical pipe section with ∆𝑧 1 𝑚, i.e. the change
in specific potential energy is of order 10 m2/s2 (9.81 m2/s2). We will also need some reference values on
variation of pressure and temperature along the conduit. Table 8‐1 shows some reference gradients of
pressure and temperature that will be use as upper limit in the derivations.

177
 Heat Transfer for Flow in Conduits M. Stanko

TABLE 8‐1. UPPER BOUNDS ON TEMPERATURE AND PRESSURE SPATIAL GRADIENTS.

Quantity Value Reference
p/L ‐7848 Pa/m Stagnant column of water
T/L ‐0.03 °C/m Geothermal gradient

8.1.1. COMPARISON BETWEEN THE SPECIFIC KINETIC AND POTENTIAL ENERGY TERMS
The change in specific kinetic energy can be neglected when it is equal to a fraction “x” (usually small) of the
change in specific potential energy:

𝑉 𝑉 EQ. 8‐6
𝑥 ∙ ∆𝑧 ∙ 𝑔
2

For this example, fraction “x” will be set to 0.1 (10%). Thus, the left side of the equation is 0.981 m2/s2.
We now express the velocity downstream in terms of the velocity upstream and a factor F:

𝑉 𝐹∙𝑉 EQ. 8‐7

Then

𝐹∙𝑉 𝑉 𝑉 EQ. 8‐8

𝑥 ∙ ∆𝑧 ∙ 𝑔 ∙ 𝐹 1
2 √2

The change in specific kinetic energy term depends on the inlet velocity of the fluid and the factor F. Figure
8‐1 shows the required values of F to equal a change in specific potential energy of 0.981 m2/s2 for different
values of 𝑉 .
1.045
1.040
1.035
1.030
1.025
F, [‐]

1.020
1.015
1.010
1.005
1.000
0 10 20 30 40 50
Vin, [m/s]

FIGURE 8‐1. PLOT OF FACTOR F VERSUS INLET FLUID VELOCITY THAT GIVES A CHANGE IN SPECIFIC KINETIC ENERGY EQUAL TO
0.981 M2/S2

Higher values of 𝑉 require lower variations between upstream and downstream velocity. In liquid flow,
velocities in the conduit are usually below 5 m/s. For the specific kinetic energy term to be negligible when
compared against the specific potential energy, the variation of velocity in one meter must be of around 3.8%
or lower.
In gas flow, velocities in the conduit are usually below 30 m/s. For the specific kinetic energy term to be
negligible when compared against the change in specific potential energy, the variation of velocity in one
meter of pipe must be of around 0.1% or lower.

178
 Heat Transfer for Flow in Conduits M. Stanko

For steady‐state conditions, the variation of velocity along the conduit is usually because of variations in
density and phase change (condensation/vaporization). Neglecting phase change, and using mass
conservation in a section of the conduit gives:

𝑉 ∙𝐴∙𝜌 𝑉 ∙𝐴∙𝜌 EQ. 8‐9

Expressing the velocity downstream in terms of the velocity upstream (with the factor F) and simplifying terms
gives:

𝜌 𝐹∙𝜌 EQ. 8‐10

This expression indicates that variations in downstream velocity are inversely proportional to variations in
downstream density. For example, if velocity increases by 30%, then downstream density must have decreased
by 30%. Variations in density are usually due to variations in pressure and temperature.
Using the values from the previous example (liquid velocity of 5 m/s), the change in specific kinetic energy will
be negligible with respect to the change in specific potential energy if there is a reduction in density of 3.8%
or lower, in a length of 1 m. In liquids, large variations in pressure are required to produce small variations in
density. Consider the formula for compressibility:

1 ∂𝜌 EQ. 8‐11
𝛽 ∙
𝜌 ∂𝑝

Approximating the differentials by deltas, gives:

1 ∆𝜌 ∆𝜌 1 1 𝜀 EQ. 8‐12
𝛽 ∙ ∙ 𝜀∙ → ∆𝑝
𝜌 ∆𝑝 𝜌 ∆𝑝 ∆𝑝 𝛽

To produce variation in density of 3.8 % (𝜀 = 0.038) in 1 meter, the variation in pressure for kerosene23 must
be of around 500 bar. This is significantly higher than pressure drops typically encountered in petroleum
production systems (usually below or slightly above the pressure gradient of a stagnant column of water, ‐
7848 Pa/m).
Therefore, for liquids, and liquid‐dominated flows, the specific kinetic energy is usually small when compared
to the potential energy and can safely be neglected.
For gases, variations of pressure and temperature have a bigger effect on density. Additionally, the density
variations required to achieve the same value of specific potential energy are smaller, because the inlet
velocity is higher.
For gases, the density at the inlet can be expressed by the real gas equation:

𝑅 EQ. 8‐13
𝑝 𝜌 ∙𝑍 ∙ ∙𝑇
𝑀

And the density at the outlet

23
𝛽=0.8 E‐5 bar‐1

179
 Heat Transfer for Flow in Conduits M. Stanko

𝑅 EQ. 8‐14
𝑝 𝜌 ∙𝑍 ∙ ∙𝑇
𝑀

If variations in temperature and gas deviation factor (Z) are considered small when compared to pressure, and
dividing both equations, this gives:

𝑝 𝜌 EQ. 8‐15
𝐹
𝑝 𝜌

Or, equivalently,

𝑝 𝐹∙𝑝 EQ. 8‐16

The pressure drop along the pipe section can then be expressed as a function of F

𝑝 1 𝐹
∆𝑝 𝑝 𝑝 𝑝 𝑝 ∙ EQ. 8‐17
𝐹 𝐹
∆𝐿 ∆𝐿 ∆𝐿 ∆𝐿

Figure 8‐2 shows values of pressure gradient versus inlet pressure that give a change in specific kinetic energy
equal to a change of specific potential energy of 0.981 m2/s2 for an inlet velocity of 30 m/s. At higher inlet
pressure, the pressure gradient required is of bigger magnitude. Values of pressure gradient above the black
line will cause the change in specific kinetic energy to be negligible.
0.0E+00
‐5.0E+03 0 100 200 300 400
pressure gradient, p/L, [Pa/m]

‐1.0E+04
‐1.5E+04
‐2.0E+04
‐2.5E+04
‐3.0E+04
‐3.5E+04
‐4.0E+04
gradient for Vin= 30 m/s
‐4.5E+04 stagnant water gradient
‐5.0E+04
pressure, p, [bara]

FIGURE 8‐2. PLOT OF PRESSURE GRADIENT VERSUS INLET PRESSURE THAT GIVE A CHANGE IN SPECIFIC POTENTIAL ENERGY EQUAL
TO 0.981 M2/S2, FOR VIN=30 M/S

Consider the pressure gradient of a stagnant column of water: ‐7848 Pa/m (red line in Figure 8‐2). Pressure
drops in single and multiphase flow in conduits of petroleum production systems are usually above or slightly
below this value. For most of the range of inlet pressure values (i.e. above 75 bara), values of pressure gradient
(for the change in specific kinetic energy to be negligible with respect to changes in specific potential energy)
are above it.
Figure 8‐3 shows the values of inlet velocity required versus inlet pressure that cause the change in specific
kinetic energy to be negligible when compared against the change in potential energy, for several values of
pressure gradients. Lower pressure gradients require higher inlet velocities for the kinetic energy term to be
negligible. For values in the upper part of the lines, the change in specific kinetic energy must likely be included.

180
 Heat Transfer for Flow in Conduits M. Stanko

200
dp/dx=‐7848 Pa/m
180 dp/dx=‐5000 Pa/m
dp/dx=‐3000 Pa/m
160
dp/dx=‐1000 Pa/m
140
120

Vin, [m/s]
100
80
60
40
20
0
0 100 200 300 400
pin, [bara]

FIGURE 8‐3. PLOT OF VALUES OF INLET VELOCITY VERSUS INLET PRESSURE THAT GIVE A CHANGE IN SPECIFIC KINETIC ENERGY
EQUAL TO 0.981 M2/S2, FOR SEVERAL VALUES OF PRESSURE GRADIENT

Therefore, for gas and gas‐dominated flow and operating conditions of low pressures and high velocities, the
term of specific kinetic energy usually must be accounted for.

8.1.2. COMPARISON BETWEEN THE SPECIFIC ENTHALPY AND POTENTIAL ENERGY TERMS
An order of magnitude comparison between the specific enthalpy and potential energy terms is expressed
mathematically as follows

𝑂 ∆𝑧 ∙ 𝑔 𝑂 ∆ℎ EQ. 8‐18

The enthalpy difference is a function of the inlet and outlet pressure and temperature and the type of fluid. In
this section, the outlet pressure and temperature will be expressed in terms of the inlet pressure and
temperature and the temperature (T/L) and pressure (p/L) gradients.
For liquids, the following approximation for enthalpy is often used:

∆ℎ 𝐶 ∙ ∆𝑇 EQ. 8‐19

𝐶 usually falls within the range between 1000‐4000 J/kg K (oil‐water).

Consider the reference vertical pipe section with ∆𝑧 1 𝑚 (change in specific potential energy equal to 9.81
m2/s2). For the change in specific enthalpy to be of the same order of magnitude the temperature gradient
T/L must equal or below values in the range ‐0.01 to ‐0.002 C/m. For the change in specific enthalpy to be
equal to 98.1 m2/s2 (10 times larger than 9.81 m2/s2), the temperature gradient must be equal or less than
values in the range ‐0.1 to ‐0.03 C/m.
In wells and pipelines, the temperature gradient along the pipe is usually higher than ‐0.03 °C/m (geothermal
temperature gradient). Therefore, for liquids and liquid‐dominated flows, it is often necessary to include the
term of change in specific potential energy, especially for liquids with low heat capacity values and when the
temperature gradient is low.
For gases, enthalpy ℎ is affected by both temperature and pressure. For methane, in the range of pressure
and temperature between 1 to 400 bara and 15 to 200 °C, the partial derivatives of enthalpy with respect to
temperature and pressure are in the ranges 2225 to 3560 J/kg/C and between ‐1.2e‐2 and ‐6.4e‐5 J/kg/Pa

181
 Heat Transfer for Flow in Conduits M. Stanko

respectively. Note that the gradients with respect to pressure and temperature have different signs. So, for
flow in pipe where both pressure and temperature are reduced, their effect will usually counteract.
Consider the reference vertical pipe section with ∆𝑧 1 𝑚 (change in specific potential energy equal to 9.81
m2/s2). For the change in specific enthalpy to be equal to 98.1 m2/s2 (10 times larger than 9.81 m2/s2), the
pressure gradient (keeping temperature constant) must be equal or lower than values between ‐1.5e‐6 and
‐8077 Pa/m. When comparing against a reference pressure gradient of stagnant water (‐7848 Pa/m) these
values are quite low.
For the change in specific enthalpy to be equal or higher than 98.1 m2/s2 (10 times larger than 9.81 m2/s2), the
temperature gradient (keeping pressure constant) must be equal or lower than values in the range ‐0.03 and
‐0.04 C/m. When comparing against a reference temperature gradient of geothermal gradient (‐0.03 C/m)
these values are quite low.
Therefore, for gases and gas‐dominated flows, it is often also necessary to include the term of change in
specific potential energy, especially for liquids with low heat capacity values and when the temperature
gradient is low.
Please note that the examples presented above were made for a vertical pipe section of 1 m. However, if the
pipe is inclined, the term of specific potential energy will be smaller (due to scaling ∆𝑒 𝑠𝑖𝑛 𝛼 ∙ ∆𝑧 ∙ 𝑔, with
𝛼 being the angle between the pipe and the horizontal) and results and limit values presented in the discussion
will be different.

8.2. HEAT TRANSFER WITH THE ENVIRONMENT

In a conduit there are usually several heat transfer mechanisms between the fluid and the ambient. For
example, there is forced convection between the fluid and the inner pipe wall, conduction in the pipe wall,
cement, insulation layer or soil, free convection in liquid trapped in the annulus, or with the external air or
water. There is also radiation in the annular space, but this is often neglected.
The heat is usually expressed in terms of an overall heat transfer coefficient, a pipe surface area, and the
temperature differential between the ambient and the fluid:

𝑄 2∙𝜋∙𝐿∙𝑟∙𝑈∙ 𝑇 𝑇 EQ. 8‐20

Where:
𝑟 reference radius [m]
𝑈 Overall heat transfer coefficient, expressed in terms of the reference radius r [W/m2.K]
𝑇 Mean ambient temperature [K or °C]
𝑇 Mean fluid temperature in the section [K or °C]

In this section we will work with the heat by unit of conduit length .

In Eq. 8‐20 the sign if positive if the temperature of the ambient is greater than the fluid while negative if
otherwise. This is consistent with the sign convention for heat (heat entering the system is positive and heat
leaving the system is negative).
The overall heat transfer coefficient U is a function of the reference radius and the number of layers and heat
transfer mechanisms between the fluid and the ambient. To explain the calculation procedure, we will
consider two cases:

182
 Heat Transfer for Flow in Conduits M. Stanko

Case 1. A subsea pipeline with four layers between the fluid and the ambient (from innermost to outer most):
 Forced convection inside the pipe
 Conduction in pipe wall
 Conduction in insulation layer
 Free/forced convection with surrounding air (or water)

Case 2. A wellbore with 7 layers between the fluid and the formation sand face (from innermost to outer
most):

 Forced convection inside the tubing

 Conduction in the tubing wall
 Free convection in the annulus between the tubing and the production casing
 Conduction in the production casing wall
 Conduction in the cement between the production and intermediate casing
 Conduction in the intermediate casing wall
 Conduction in the cement between the intermediate casing and the formation

8.2.1. CASE 1. SUBSEA PIPELINE

A sketch of the configuration to study is shown in Figure 8‐4.

FIGURE 8‐4. SKETCH SHOWING THE SIDE VIEW OF A SUBSEA PIPELINE WITH INNER FORCED CONVECTION, CONDUCTION IN PIPE
WALL AND INSULATION AND FREE CONVECTION WITH WATER

We will first write the expressions to calculate heat transfer in each layer.
Forced convection inside the pipe
The heat transfer in the interior of the flowing pipe can be expressed as follows:

𝑑𝑄 EQ. 8‐21
2∙𝜋∙𝑟 ∙ℎ ∙ 𝑇 𝑇
𝑑𝐿

183
 Heat Transfer for Flow in Conduits M. Stanko

Where:
𝑟 Pipe inner radius [m]
ℎ Convective coefficient, inner fluid [W/m2.K]
𝑇 Fluid temperature [K or °C]
𝑇 Temperature, inner pipe wall [K or °C]

The convective heat transfer coefficient ℎ , depends on the fluid and wall temperatures, velocities and
distribution of phases inside the pipe, variables which are the result of pressure and temperature drop
calculations. It is often calculated from the Nusselt number (ratio of convective and conductive heat transfer
h∙L /k). The Nusselt number is often expressed as a function of the Reynolds number and the Prandtl number
(μ∙Cp / k).
Therefore, the estimation of ℎ is often an implicit calculation (because fluid and wall temperature and
velocities are not known a priori). Initially a value is assumed, pressure and temperature are calculated in the
conduit, and then a new value of ℎ is estimated. The process is repeated iteratively until convergence is
achieved.
Conduction in pipe wall
The heat transfer in the metal wall of the pipe can be expressed with the following expression:

𝑑𝑄 𝑇 𝑇 EQ. 8‐22
2∙𝜋∙𝑘 ∙ 𝑟
𝑑𝐿 ln
𝑟
Where:
𝑟 Pipe inner radius [m]
𝑟 Pipe outer radius [m]
𝑘 Pipe material thermal conductivity [W/m.K]
𝑇 temperature, inner pipe wall [K or °C]
𝑇 temperature, outer pipe wall [K or °C]

Conduction in insulating layer

Similar to the heat transfer in the metal wall of the pipe, the heat transfer in the insulation layer can be
expressed with the following expression:

𝑑𝑄 𝑇 𝑇 EQ. 8‐23
,
2∙𝜋∙𝑘 ∙ 𝑟
𝑑𝐿 ln ,
𝑟
Where:
𝑟 , Insulation outer radius [m]
𝑟 Pipe outer radius [m]
𝑘 Insulation material thermal conductivity [W/m.K]
𝑇 Temperature, outer pipe wall [K or °C]
𝑇 , Temperature, outer insulation wall [K or °C]

184
 Heat Transfer for Flow in Conduits M. Stanko

Free‐forced convection with air (or water)

The heat transfer (free and forced convection) with external air (or water) can be expressed with the following
expression:

𝑑𝑄 EQ. 8‐24
2∙𝜋∙𝑟 , ∙ℎ ∙ 𝑇 , 𝑇
𝑑𝐿
Where:
𝑟 , Insulation outer radius [m]
ℎ Convective coefficient, outer fluid [W/m2.K]
𝑇 Ambient temperature (sea water) [K or °C]
𝑇 , Temperature, outer insulation wall [K or °C]

Similar to the heat transfer inside the pipe, this heat transfer mechanism is usually a combination of free
convection (the flow is induced by the temperature difference) and forced convection (there is an external
current e.g. wind or marine current). The convective heat transfer coefficient ℎ depends on the fluid and wall
temperatures, velocities and distribution of phases inside the pipe, variables which are the result of pressure
and temperature drop calculations and are not known a priori. It is often calculated from the Nusselt number
(ratio of convective and conductive heat transfer h∙L /k). The Nusselt number is often expressed as a function
of the Reynolds number, the Prandtl number (μ∙Cp /k) and the Grashoff number (g∙ρ2∙|ΔT|∙L3∙|β| /μ2). β is the
thermal volumetric expansion coefficient at constant temperature, equal to ∙

Therefore, the estimation of ℎ often requires an implicit calculation where a value is assumed, pressure and
temperature are calculated in the conduit, and then a new value of ℎ is estimated. The process is repeated
iteratively until convergence is achieved.
Overall heat transfer coefficient
The overall heat transfer coefficient is estimated by clearing temperature difference and summing up all
expressions (Eq. 8‐21, Eq. 8‐22, Eq. 8‐23, and Eq. 8‐24):

𝑇 𝑇 𝑇 𝑇 𝑇 𝑇 , 𝑇 𝑇 EQ. 8‐25
,
𝑑𝑄 𝑑𝑄 𝑑𝑄 𝑑𝑄
𝑑𝐿 𝑑𝐿 𝑑𝐿 𝑑𝐿
2∙𝜋∙𝑟 ∙ℎ 2∙𝜋∙𝑘 2∙𝜋∙𝑘 2∙𝜋∙𝑟 , ∙ℎ
𝑟 𝑟 ,
ln ln
𝑟 𝑟

Clearing the temperature difference between fluid and environment:

EQ. 8‐26
⎡ ⎤
𝑑𝑄 ⎢ 1 1 1 1 ⎥
𝑇 𝑇 ∙⎢ ⎥
𝑑𝐿 ⎢2 ∙ 𝜋 ∙ 𝑟 ∙ ℎ 2∙𝜋∙𝑘 2∙𝜋∙𝑘 2∙𝜋∙𝑟 , ∙ℎ ⎥
⎢ 𝑟 𝑟 , ⎥
ln ln
⎣ 𝑟 𝑟 ⎦
If the inner pipe radius will be used as reference radius, we then divide Eq. 8‐26 by the inner pipe perimeter:

185
 Heat Transfer for Flow in Conduits M. Stanko

EQ. 8‐27
⎡ ⎤
𝑑𝑄 1 ⎢1 1 1 1 ⎥
𝑇 𝑇 ∙ ∙⎢ 𝑟 , ⎥
𝑑𝐿 2 ∙ 𝜋 ∙ 𝑟 ⎢ℎ 𝑘 𝑘 ∙ℎ ⎥
⎢ 𝑟 𝑟 , 𝑟 ⎥
𝑟 ∙ ln 𝑟 ∙ ln
⎣ 𝑟 𝑟 ⎦
The overall heat transfer coefficient based on the pipe inner area is defined as:

𝑟 𝑟 ,
EQ. 8‐28
1 𝑟 ∙ ln 𝑟 ∙ ln 𝑟
𝑟 𝑟
𝑈
ℎ 𝑘 𝑘 𝑟 , ∙ℎ

There are some cases where a few terms in this expression are much larger than others, thus a few can be
safely neglected:
 Inner forced convection: The inner forced‐convection coefficient (ℎ ) is usually in the range 100‐
50 000 W/m2 K.24 It is lower for low velocities and for gas flow. This gives a term in the range O(1E‐
5) to O(1E‐2).
 Conduction in metal: Inner radii of well tubulars and pipelines are usually in the range 0.01‐0.25 m.
The ratio between inner and outer radius is usually between 1.05‐1.3 (thickest pipe walls are usually
for the small pipe diameters), thus the natural log of it is between 0.04‐0.24. Lastly, the conductivity
of the steel is around 45 W/m2 K. This gives a term O(1E‐4).
 Conduction in the insulating layer: Inner radii of well tubulars and pipelines are usually in the range
0.01‐0.25 m. The ratio between inner and outer radius of the insulating layer is usually around 2,
thus the natural log is around 0.6. The conductivity of the insulating layer is usually around 0.2‐0.3
W/m2 K (polypropylene). This gives a term in the range O(1E‐2) to O(1).
 Free/forced outer convection: The ratio between the inner pipe wall and the outermost radius is
more than 2. The free convection coefficient is usually a number between 1‐400 W/m2 K (for air, it is
lower, and for water it is higher, and for moving fluids it is higher, while for stagnant fluids it is
lower). This gives a term between O(1E‐3) and O(1E‐1).

In this case, the most important contribution to the overall heat transfer coefficient comes from the
conduction in the insulation layer and (maybe) inner and outer convection. This is important to consider when
modeling temperature and pressure drop of multiphase flow in wellbores, flowlines and pipelines, to avoid
performing implicit calculations. If it is found that the inner or outer convection coefficient will not contribute
significantly to the overall heat transfer coefficient, then can often be neglected and the calculation executed
explicitly thus avoiding iteration.
If the outer area is used instead of the inner area, then the overall heat transfer coefficient based on the outer
area is the following:

𝑟 𝑟 ,
EQ. 8‐29
𝑟 , 𝑟 , ∙ ln 𝑟 , ∙ ln 1
𝑟 𝑟
𝑈
𝑟 ∙ℎ 𝑘 𝑘 ℎ

The relationship between the overall heat transfer coefficient based on inner area (Ui) and the overall heat
transfer coefficient based on outer area (Uo) is :

24
https://www.engineeringtoolbox.com/convective‐heat‐transfer‐d_430.html

186
 Heat Transfer for Flow in Conduits M. Stanko

𝑈 𝑟 EQ. 8‐30
𝑈 𝑟 ,

8.2.1. CASE 2. HEAT TRANSFER IN WELLBORE

A sketch of the configuration to study is shown in Figure 8‐5.

FIGURE 8‐5. SKETCH SHOWING THE SIDE VIEW OF A WELLBORE WITH INNER FORCED CONVECTION, CONDUCTION IN TUBING
WALL, FREE CONVECTION IN ANNULUS, CONDUCTION IN WALL OF PRODUCTION CASING, CONDUCTION IN CEMENT LAYER
BETWEEN PRODUCTION CASING AND INTERMEDIATE CASING, CONDUCTION IN WALL OF INTERMEDIATE CASING, CONDUCTION IN
CEMENT LAYER BETWEEN FORMATION AND INTERMEDIATE CASING

Where:

𝑟, Tubing inner radius [m]

𝑟, Tubing outer radius [m]
𝑟 , Production casing inner radius [m]
𝑟 , Production casing outer radius [m]
𝑟 , Intermediate casing inner radius [m]
𝑟 , Intermediate casing outer radius [m]
𝑟 Wellbore radius [m]
𝑇, Temperature at tubing inner radius [m]
𝑇, Temperature at tubing outer radius [m]

187
 Heat Transfer for Flow in Conduits M. Stanko

𝑇 , Temperature at production casing inner radius [m]

𝑇 , Temperature at production casing outer radius [m]
𝑇 , Temperature at intermediate casing inner radius [m]
𝑇 , Temperature at intermediate casing outer radius [m]
𝑇 Temperature at wellbore radius [m]

We will first write the expressions to calculate heat transfer in each layer. The expressions for forced
convection inside the tubing, conduction in the tubing wall, in the production casing wall, in the cement
between the intermediate and production casing and in the cement between the intermediate casing and
formation are similar to the ones presented earlier for the case of the subsea pipeline and won’t be repeated.
Free convection in the annulus
The heat transfer (free convection) in the annulus fluid can be expressed with the following expression:

𝑑𝑄 EQ. 8‐31
2∙𝜋∙𝑟, ∙ℎ ∙ 𝑇, 𝑇 ,
𝑑𝐿
Where:
𝑟, Tubing outer radius [m]
ℎ Free convection coefficient, annulus [W/m2.K]
𝑇, Temperature of outer wall of tubing [K or °C]
𝑇 , Temperature of inner wall of intermediate casing [ or °C]

This heat transfer mechanism is usually free convection (the flow is induced by the temperature difference).
The convective heat transfer coefficient ℎ depends on the fluid and wall temperatures, variables which are
the result of pressure and temperature drop calculations and are not known a priori. It is often calculated from
the Nusselt number (ratio of convective and conductive heat transfer h∙L /k). The Nusselt number is often
expressed as a function of the Reynolds number, the Prandtl number (μ∙Cp /k) and the Grashoff number
(g∙ρ2∙|ΔT|∙L3∙|β| /μ2). β is the thermal volumetric expansion coefficient at constant temperature, equal to
∙

Therefore, the estimation of ℎ often requires an implicit calculation where a value is assumed, pressure
and temperature are calculated in the conduit, and then a new value of ℎ is estimated. The process is
repeated iteratively until convergence is achieved.
Overall heat transfer coefficient
The overall heat transfer coefficient is estimated by clearing temperature difference and summing up all
expressions, as done previously:

188
 Heat Transfer for Flow in Conduits M. Stanko

𝑇 𝑇, 𝑇, 𝑇, 𝑇, 𝑇 , 𝑇 , 𝑇 , 𝑇 , 𝑇 , 𝑇 , 𝑇 ,
𝑇 , 𝑇
𝑑𝑄 𝑑𝑄 𝑑𝑄 𝑑𝑄 𝑑𝑄
𝑑𝐿 𝑑𝐿 𝑑𝐿 𝑑𝐿 𝑑𝐿
2∙𝜋∙𝑟, ∙ℎ 2∙𝜋∙𝑘 2∙𝜋∙𝑟, ∙ℎ 2∙𝜋∙𝑘 2∙𝜋∙𝑘
𝑟 𝑟 , 𝑟 , EQ. 8‐32
ln , ln ln
𝑟, 𝑟 , 𝑟 ,
𝑑𝑄 𝑑𝑄
𝑑𝐿 𝑑𝐿
2∙𝜋∙𝑘 2∙𝜋∙𝑘
𝑟 , 𝑟
ln ln
𝑟 , 𝑟 ,

Clearing the temperature difference between fluid and wellbore wall:

𝑑𝑄
𝑇 𝑇
𝑑𝐿
⎡
⎢ 1 1 1 1 1
∙⎢
⎢ 2 ∙ 𝜋 ∙ 𝑟, ∙ℎ 2∙𝜋∙𝑘 2∙𝜋∙𝑟, ∙ℎ 2∙𝜋∙𝑘 2∙𝜋∙𝑘
⎢ 𝑟 𝑟 , 𝑟 , EQ. 8‐33
ln , ln ln
⎣ 𝑟, 𝑟 , 𝑟 ,

⎤
1 1 ⎥
⎥
2∙𝜋∙𝑘 2∙𝜋∙𝑘 ⎥
𝑟 , 𝑟 ⎥
ln ln
𝑟 , 𝑟 , ⎦

If the inner tubing radius will be used as reference radius, we then we divide by the inner perimeter of the
inner tubing:

𝑑𝑄 1
𝑇 𝑇 ∙
𝑑𝐿 2 ∙ 𝜋 ∙ 𝑟 ,
𝑟, 𝑟 , 𝑟 ,
𝑟 , ∙ ln 𝑟 , ∙ ln 𝑟 , ∙ ln
1 𝑟, 𝑟, 𝑟 , 𝑟 ,
∙
ℎ 𝑘 𝑟, ∙ℎ 𝑘 𝑘
EQ. 8‐34
𝑟 , 𝑟
𝑟 , ∙ ln 𝑟 , ∙ ln
𝑟 , 𝑟 ,
𝑘 𝑘

The overall heat transfer coefficient expressed in terms of the inner radius of the tubing is:
𝑟, 𝑟 , 𝑟 , 𝑟 , EQ. 8‐35
𝑟 , ∙ ln 𝑟 , ∙ ln 𝑟 , ∙ ln 𝑟 , ∙ ln
1 𝑟, 𝑟, 𝑟 , 𝑟 , 𝑟 ,
𝑈
ℎ 𝑘 𝑟, ∙ℎ 𝑘 𝑘 𝑘

𝑟
𝑟 , ∙ ln
𝑟 ,
𝑘

Where:

189
 Heat Transfer for Flow in Conduits M. Stanko

ℎ Convective coefficient, inner fluid [W/m2.K]

𝑘 Thermal conductivity of tubing [W/m.K]
ℎ Free convection coefficient, annulus [W/m2.K]
𝑘 Thermal conductivity of cement [W/m.K]
𝑘 Thermal conductivity of production casing [W/m.K]
𝑘 Thermal conductivity of intermediate casing [W/m.K]

As mentioned earlier for the case of the subsea pipeline, the terms for metal conduction (tubing and casing
wall) are usually small and can be neglected (terms 2, 4 and 6). The term of forced convection in the pipe might
be in the range O(1E‐5) to O(1E‐2) and could be neglected.
Free convection in the annulus (Term 3): The free convection coefficient in the annulus usually has values
around 100 W/m2 K. The ratio between outer and inner tubing diameter can range from 1.05 to 1.3. Therefore,
this term is usually O(1E‐2).
Conduction in cement (terms 5 and 7): The thermal conductivity of cement (𝑘 ) is usually in the range between
0.3 to 2 W/m K. The ratio between the outer and inner diameter of the annular space is usually around 1.2.
The inner tubing diameter is usually 0.02‐0.2. Therefore, this term is usually O(1E‐2).
Heat transfer in formation or soil
In the previous derivation, we were assuming that the temperature of the sandface of the formation is
constant. However, this is often not the case when the well is producing or injecting. In these situations, the
temperature of the formation near the wellbore will change (will be warmed up or cooled down) and the
undisturbed formation temperature (geothermal) will be located farther away from the wellbore. For large
times, a steady state will be reached where the temperature distribution in the soil will not change any longer.
The heat transfer in the formation is a transient problem and is often described with the following partial
differential equation (here Te is soil temperature and r is radius):
EQ. 8‐36
∂ 𝑇 1 ∂𝑇 𝑐 ∙ 𝜌 ∂𝑇
∙ ∙
∂𝑟 𝑟 ∂𝑟 𝑘 ∂𝑡

Where:

𝑘 Thermal conductivity, soil [W/m.K]

𝐶 Specific heat capacity, soil [J/K.kg]
𝑡 Time [s]

This equation is made for a given vertical position and is considering the heat transfer is mostly radial.
This equation has often the following initial conditions and boundary conditions:
EQ. 8‐37
𝑇 𝑟, 𝑡 0 𝑇

∂𝑇 EQ. 8‐38
𝑟 → ∞, 𝑡 0
∂𝑟

190
 Heat Transfer for Flow in Conduits M. Stanko

EQ. 8‐39
𝑑𝑄 ∂𝑇
2∙𝜋∙𝑘 ∙𝑟 ∙
𝑑𝑧 ∂𝑟

The last boundary condition represents the heat transfer (per unit length) that is transferred with the well.
Several exact and approximate solutions have been proposed to this equation. An approach consists of
defining a dimensionless temperature:

2∙𝜋∙𝑘 EQ. 8‐40

𝑇 𝑇 𝑇
𝑑𝑄
𝑑𝑧

The behavior of dimensionless temperature with time can be approximated with the following expressions[8‐
1]
:
EQ. 8‐41
𝑇 1.1281 ∙ 𝑡 ∙ 1 0.3 ∙ 𝑡 , 𝑓𝑜𝑟 𝑡 1.5

0.6 EQ. 8‐42

𝑇 0.4063 0.5 ∙ 𝑙𝑛 𝑡 1 𝑓𝑜𝑟 𝑡 1.5
𝑡

With:
∙
Dimensionless time, 𝑡

And 𝛼 is the thermal diffusivity of the soil, [m2/s], equal to

∙

The solution above provides an expression of the temperature of the sandface (𝑇 ) with time, if the heat rate
per unit length ( ) is provided.

Now we will include the effect of the soil transient in the wellbore heat transfer equation. The heat transfer
equation considering all heat transfer mechanisms from fluid to wellbore (sandface) is:
EQ. 8‐43
𝑑𝑄
2∙𝜋∙𝑟, ∙𝑈∙ 𝑇 𝑇
𝑑𝐿

Clearing out the temperature of the sandface (Twb):

EQ. 8‐44
𝑑𝑄 1
𝑇 𝑇 ∙
𝑑𝐿 2 ∙ 𝜋 ∙ 𝑟 , ∙ 𝑈

This expression is now substituted in the definition of dimensionless temperature, TD:

EQ. 8‐45
2∙𝜋∙𝑘 𝑑𝑄 1
𝑇 𝑇 ∙ 𝑇
𝑑𝑄 𝑑𝐿 2 ∙ 𝜋 ∙ 𝑟 , ∙ 𝑈
𝑑𝑧

The wellbore usually has some inclination, and therefore the vertical and along the pipe coordinates do not
overlap. However, it is often assumed that this difference can be neglected and the terms and have the

same magnitude. However, they have different signs ( ). For the wellbore, if the fluid is hotter than

191
 Heat Transfer for Flow in Conduits M. Stanko

the formation, the heat is lost to the formation, therefore the sign is negative. For the formation, if the fluid is
hotter than the formation, the heat enters the formation, so the sign is positive.
Simplifying terms and rearranging:

𝑘 2∙𝜋∙𝑘 EQ. 8‐46

𝑇 𝑇 𝑇
𝑟, ∙𝑈 𝑑𝑄
𝑑𝐿

Clearing out the heat rate per unit length

EQ. 8‐47
𝑑𝑄 2∙𝜋∙𝑘
𝑇 𝑇
𝑑𝐿 𝑘
𝑇
𝑟, ∙𝑈

Separating terms to obtain the inner perimeter of tubing

EQ. 8‐48
𝑑𝑄 𝑈∙𝑘
2∙𝜋∙𝑟, ∙ 𝑇 𝑇
𝑑𝐿 𝑘 𝑇 ∙𝑟, ∙𝑈

This equation shows that the transient in the formation causes the effective overall heat transfer coefficient
to be modified according to the following expression:
EQ. 8‐49
𝑈∙𝑘
𝑈 𝑡
𝑘 𝑇 ∙𝑟, ∙𝑈

The thermal conductivity of the formation (𝑘 is usually a number around 3 W/mK. Dimensionless
temperature (𝑇 ) is usually a number between 0 (early times) to 7 (late times). The inner tubing radius is
usually ranging from 0.05 to 0.23 m. Overall heat transfer coefficient for wellbores are usually in the range 3‐
15 W/m2 K. Therefore, for late times, the effective overall heat transfer coefficient is usually lower than the
overall heat transfer coefficient neglecting the formation. This means that the hot formation region around
the wellbore with time acts as an insulating layer.

8.3. BEHAVIOR OF SPECIFIC ENTHALPY OF OIL AND GAS VERSUS PRESSURE AND TEMPERATURE FOR
MULTI‐COMPONENT HYDROCARBON MIXTURES

Figure 8‐6 presents a sketch of the variation of the oil and gas specific enthalpy properties versus pressure for
three temperatures of the hydrocarbon mixture shown in Figure 8‐7. At the given temperatures, single phase
oil will be formed for pressures equal or greater than the bubble point pressures.

192
 Heat Transfer for Flow in Conduits M. Stanko

OIL PHASE GAS PHASE

FIGURE 8‐6. BEHAVIOR OF SPECIFIC ENTHALPY OF GAS AND OIL VS. PRESSURE FOR THREE TEMPERATURES

FIGURE 8‐7. PHASE DIAGRAM OF THE HYDROCARBON MIXTURE USED IN FIGURE 8‐6

Figure 8‐8 presents a sketch of the variation of the oil and gas specific enthalpy properties versus pressure for
three temperatures of the hydrocarbon mixture shown in Figure 8‐9. At the given temperatures, single phase
gas will be formed for pressures equal or greater than the dew point pressures.

OIL PHASE GAS PHASE

FIGURE 8‐8. BEHAVIOR OF SPECIFIC ENTHALPY OF GAS AND OIL VS. PRESSURE FOR THREE TEMPERATURES

193
 Heat Transfer for Flow in Conduits M. Stanko

FIGURE 8‐9. PHASE DIAGRAM OF THE HYDROCARBON MIXTURE USED IN FIGURE 8‐8

The behavior shown in the figure corresponds to multi‐component mixtures. For fluids consisting of pure
components, the behavior in the single‐phase region (liquid or gas) is similar, but in the saturation region it
isn’t, since there is only one saturation temperature (Tbubblepoint =Tdewpoint) at a given pressure. Therefore, in
saturation, a pressure‐temperature combination could give you several values of specific enthalpy.

8.4. PROCEDURE TO ESTIMATE TEMPERATURE DROP IN A CONDUIT

Split the conduit in intervals and start your computations from a boundary point of known temperature. For
every interval in the conduit, the temperature at the other end of the interval can be computed as follows:
 The heat exchanged by the section is computed using the temperature difference between the known
temperature at one end of the interval (𝑇 or 𝑇 ) and the temperature of the environment (𝑇 )
at the corresponding end. The calculation requires the overall heat transfer coefficient of the section
(U), the pipe inner diameter (𝜙) and the length of the interval (L).
𝑄 𝑇 𝑇 @ ∙ 𝑈 ∙ 𝜋 ∙ 𝜙 ∙ ∆𝐿 EQ. 8‐50
 The enthalpy at the other end of the interval is calculated with the energy equation, neglecting
changes in specific kinetic energy
∆ℎ ∆𝑧 ∙ 𝑔 ∙ 𝑚 𝑄 → ℎ ℎ ∆𝑧 ∙ 𝑔 ∙ 𝑚 𝑄 EQ. 8‐51

 The temperature at the other pipe end is calculated with the PVT model, the value of specific enthalpy
and the pressure. This usually requires some iteration to converge, e.g. assume a value of
temperature, compute enthalpy with the PVT model, the pressure and assumed temperature and
then verify that the computed enthalpy is equal than the one obtained with the energy balance.
This procedure assumes that pressure is known at the end of the segment. However, pressure calculations in
a conduit also require a temperature input. Therefore, these must often be executed simultaneously.

194
Heat Transfer for Flow in Conduits M. Stanko

REFERENCES
[8‐1] Hasan, A.R.; Kabir, C.S. Heat Transfer During Two‐Phase Flow in Wellbores: Part I – Formation
Temperature. SPE paper 22866. 66th Annual Technical Conference and Exhibition. Dallas, October,
1991.

195
 Gas well liquid loading M. Stanko

9. GAS WELL LIQUID LOADING

Liquid loading is a phenomenon that occurs in gas wells producing associated liquid (e.g. condensate, water)
whose symptoms are a sharp decline in production, intermittent production and presence of stagnant liquids
in the wellbore. A liquid loaded well will eventually be unable to flow unaided.
It is commonly accepted that liquid loading occurs when the flow pattern of part of the wellbore starts
transitioning from gas‐dominated to liquid‐dominated. Usually slug, churn and bubble are considered as liquid‐
dominated flow patterns and mist, annular as gas‐dominated.
Figure 9‐1 shows the flow pattern map of vertical upward flow in pipe for a specific set of fluid properties and
pipe diameter. Gas‐dominated flow patterns require high superficial velocities25 of gas and relatively low
superficial velocities of liquid.
Note the transitions are often not sharp as drawn in the figure, but gradual (e.g. in bands instead of lines).
Additionally, visual classification of flow regimes can be highly subjective.
10

Dispersed
superficial velocity of liquid, vsl, [m/s]

0.1
Bubble Slug
Churn
Annular
0.01

0.001
0.01 0.1 1 10 100
superficial velocity of gas, vsg, [m/s]

FIGURE 9‐1. SAMPLE FLOW PATTERN MAP FOR VERTICAL (90°) UPWARD FLOW OF GAS AND LIQUID

Transition criteria for liquid loading is often built based on determining when the liquid phase will start to flow
downwards starting from a gas‐dominated flow pattern. For example, if one assumes that most of the liquid
is transported in droplets, then liquid loading will occur when the gas is not able to drag droplets upwards and
start traveling downwards. If, on the contrary, one assumes that most of the liquid is transported as a liquid
on the wall, then liquid loading will occur when the velocity of the liquid at the film starts pointing downwards
(Figure 9‐2).

25
Local rate of gas divided by the total pipe cross section area

196
 Gas well liquid loading M. Stanko

a) b)
FIGURE 9‐2. VISUALIZATION OF LIQUID LOADING TRANSITION CRITERIA WHEN THE LIQUID IS TRANSPORTED MAINLY AS A FILM ON
THE PIPE WALL. A) REPRESENTS AN UNLOADED CONDITION (ANNULAR FLOW) AND B) REPRESENTS A LOADED CONDITION (CHURN‐
ANNULAR). TAKEN FROM VAN’T WESTENDE[7‐3]

When loading occurs, it does not imply that the flow pattern will change immediately from gas to liquid
dominated. It usually takes some further reduction in gas rate (or increase in the liquid rate) to fully achieve
the transition.
The scientific and technical community has done extensive work in the past years to develop mechanistic and
empirical models to estimate accurately the point where loading occurs. Some focus issues are how is the
liquid distributed in the pipe (in droplets entrained in the gas or transported as a film on wall, or both) and the
effect of pipe inclination on the asymmetry of the wall liquid film.
A typical procedure to determine if a well has liquid loading problems is to perform temperature and pressure
drop calculations along the wellbore for its current operating conditions and apply a liquid loading criterion at
all depths. Alternatively, to determine the flow pattern for all depths.
The onset of liquid loading is typically determined by finding the critical gas standard condition rate that causes
the wellbore to exhibit liquid loading (or a liquid‐dominated flow pattern) at least in one location.
As an example, consider a vertical gas well producing with a constant wellhead pressure of 14.8 bara, that has
a total depth of 3 000 m, and inner tubing diameter of 0.15 m. The well is producing water, has a water gas
ratio (WGR) of 5 stb/1E06 scf26 and a condensate‐gas ratio of 0 stb/1E06 scf. In this example, water is used
instead of condensate, to partly remove the complexity of change in the mass rate of condensate along the
tubing due to condensation or evaporation. The model did not consider vaporization of water in the gas.
Multiphase flow calculations are made with a mechanistic model.
Figure 9‐3 shows the calculated flow pattern along the tubing for several values of gas rate (at standard
conditions). A TVD value of 0 m is the wellhead, and a TVD value of 3000 m is the bottom‐hole. When the gas
rate is reduced (from left to right) the wellbore transitions from flow patterns that are gas dominated (annular)
to flow patterns that are liquid dominated (slug).

26
Here field units are used because they are easier to remember

197
 Gas well liquid loading M. Stanko

If one sets that the flow pattern “slug” is considered as liquid loaded, then the gas flow rate of 0.24 E06 Sm3/d
will be the critical loading gas flow rate, because that is the first rate for which a location in the tubing exhibits
loading. To have an “unloaded” well, one must produce the well above this rate.
0 Annular Annular Annular Slug

Annular Annular Annular Slug

Annular Annular Annular Slug

Annular Annular Annular Slug

500
Annular Annular Annular Slug

Annular Annular Annular Slug

Annular Annular Annular Slug

True vertical depth, TVD, [m] 1000
Annular Annular Annular Slug

Annular Annular Annular Slug

Annular Annular Annular Slug

1500 Annular Annular Slug Slug

Annular Annular Slug Slug

Annular Annular Slug Slug

Annular Annular Slug Slug

2000
Annular Annular Slug Slug

Annular Annular Slug Slug

Annular Annular Slug Slug

2500
Annular Annular Slug Slug

Annular Annular Slug Slug

Annular Annular Slug Slug

3000 Annular Slug Slug Slug

qg [1E06 Sm3/d]: 0.85 0.24 0.22 0.14

FIGURE 9‐3. FLOW PATTERNS ALONG THE TUBING FOR THE TEST WELL FOR SEVERAL VALUES OF GAS RATE

The gas rate of a gas well will usually diminish with time, due to depletion. When this happens the gas well
usually transitions from an unloaded to a loaded condition. The bottom of the well is usually the place where
liquid loading occurs first. This is because the pressure is highest, the volumetric local rate of gas is smaller,
and therefore the gas velocity (and superficial velocity) is highest.
However, the liquid rate also affects the flow pattern and liquid loading (ref. Figure 9‐1). In some gas wells,
due to condensation at low pressure and temperature, there is usually more free liquid at the wellhead than
at the bottom‐hole. Therefore, in these cases liquid loading occurs usually in the shallow part of the wellbore,
despite the high gas velocities.
Figure 9‐4 shows the flowing bottom‐hole pressure and the percentage of wellbore volume occupied by the
liquid versus standard conditions gas rates for the example case discussed earlier. The gas rate at which the
“slug” flow pattern is detected in the tubing is marked with a vertical discontinuous violet line. For values of
gas rates below the critical, the flowing bottom‐hole pressure and percent of wellbore filled by the liquid start
to increase abruptly. In this case, liquid loading is detrimental for well performance, because the pressure drop
in the tubing increases with reduction in gas rate.

198
 Gas well liquid loading M. Stanko

90 25

80

flowing bottomhole pressure, pwf, [bara]

20

% of wellbore filled with liquid

70

60
15
50

40
10
30

20 5
10

0 0
0.00 0.20 0.40 0.60 0.80 1.00
gas rate, qg, [1E06 Sm3/d]
flowing bottom‐hole pressure % of wellbore filled with liquid

FIGURE 9‐4. FLOWING BOTTOM‐HOLE PRESSURE AND PERCENTAGE OF WELLBORE VOLUME OCCUPIED BY THE LIQUID VERSUS
STANDARD CONDITIONS GAS RATE

9.1. LIQUID LOADING MAPS

Liquid loading maps are plots that display the relationship between condensate gas ratio (or gas‐oil ratio, GOR,
or gas liquid ratio, GLR) and critical loading gas rate. Figure 9‐5 shows a liquid loading map calculated for the
example discussed earlier. This liquid loading map is made for a constant wellhead pressure. The critical gas
rate is computed using the bisection method, starting with two extreme values: a very low gas rate (that will
give loaded conditions) a very high gas rate (that will give unloaded conditions). The condensate‐gas ratio is
equal to 0 stb/1E06 scf. As mentioned earlier, WGR is used instead of CGR to remove complexities due to
phase change. The model did not consider vaporization of water in the gas.
For this example, we are considering the flow pattern “slug” as liquid loaded.
45
LOADED
40
Water gas ratio, WGR, [stb/1E06 scf]

35

30
LOADED
25

20
UNLOADED
15

10

0
0.00 0.10 0.20 0.30 0.40 0.50 0.60
gas rate, qg, [1E06 Sm3/d]

FIGURE 9‐5. CURVES OF CRITICAL GAS RATE VERSUS WATER GAS RATIO FOR A WELLHEAD PRESSURE OF 14.8 BARA.

This figure shows that the critical gas rate increases with the WGR. This is because when the amount of liquid
in the tubing increases, the superficial velocity of the liquid increases, and it is more likely the gas‐liquid
superficial velocity combination will fall in a liquid‐dominated flow pattern region. However, when the WGR is
greater than 40 stb/1E06 scf, no matter what gas rate the well is produced at, the well will always be loaded.

199
 Gas well liquid loading M. Stanko

This could be explained by using at the flow map shown in Figure 9‐1. For usl > 0.2 m/s, the flow pattern will
always be slug, no matter how high the gas usg is.
Figure 9‐6 shows the curves of critical gas rate versus condensate gas ratio for 3 values of wellhead pressure.
Higher wellhead pressures have higher critical gas rates than lower wellhead pressures. This is because the
gas volumetric rate is reduced at high pressure, and therefore superficial gas velocity will also be less.
45

Water gas ratio, WGR, [stb/1E06 scf] 40

35

30

25

20

15
pwh=15 bara
10 pwh=50 bara
pwh=100 bara
5

0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
gas rate, qg, [1E06 Sm3/d]

FIGURE 9‐6. CURVES OF CRITICAL GAS RATE VERSUS WATER GAS RATIO FOR SEVERAL VALUES OF WELLHEAD PRESSURE

Figure 9‐7 shows curves of WGR and critical gas rate for 3 values of tubing inner diameter at a constant
wellhead pressure of 14.8 bara. Smaller tubing diameters move the curve to the left (require lower values of
critical gas rates). This is because the superficial gas velocity is greater for lower diameters.
40

35
Water gas ratio, WGR, [stb/1E06 scf]

30

25

20

15
ID=0.15 m
10 ID=0.10 m
ID=0.05 m
5

0
0.00 0.10 0.20 0.30 0.40 0.50 0.60
gas rate, qg, [1E06 Sm3/d]

FIGURE 9‐7. CURVES OF CRITICAL GAS RATE VERSUS WATER GAS RATIO FOR SEVERAL VALUES OF TUBING INNER DIAMETER

Liquid loading maps can be used to predict when a well will enter in liquid loading (e.g. by plotting the predicted
pairs of CGR‐qg and determine when they cross the boundary). Also, to determine what kind of changes can
be done to the well (e.g. change tubing size, install a velocity string, change the wellhead pressure by choking,
apply gas‐lift etc.) to move the operational conditions from a loading state to a unloaded state.

200
 Gas well liquid loading M. Stanko

9.2. “TRANSIENTNESS” OF LIQUID LOADING

Not so transient after all?
As shown in the previous sections, when the well gas rate is reduced (or the CGR increases) this usually leads
to an increase in the liquid content of the wellbore. For this to occur, there must a transient liquid buildup
process from a condition 1 (CGR1, qg1) to a condition 2 (CGR2, qg2) where part of the reservoir liquids will not
be produced to surface but will be gradually “retained” in the wellbore. However, such liquid volumes and
usually small and the accumulation process occurs gradually over long time (days‐weeks‐months). Therefore,
the evolution of the wellbore conditions in time can often be properly approximated by a sequence of steady
states.
Transition criteria for liquid loading is often built based on determining when the liquid phase will start to flow
downwards considering a gas‐dominated flow pattern. However, this does not mean that the net flow of liquid
will be downwards once the well is loaded (e.g. the liquid will drip and accumulate at the bottom of the
wellbore and never flows to surface). When liquid starts traveling downwards in a mist/annular flow pattern,
the flow pattern will eventually change to churn/intermittent flow, characterized by quick transients with
upwards and downwards flow of pockets of gas, slugs of liquid and liquid waves in a chaotic manner. In these
flow patterns, the net liquid and gas flux over a sufficiently long time (seconds/minutes) are still upwards, but
there are some short quick periods where there the flow is downwards or there might be part of the cross‐
section area with fluid recirculation. Intermittent flow patterns such as slug and churn can still be modeled
with steady state approximations.
When churn and slug flow occur at the bottom of the well, there will be a significant pressure gradient along
the sand face. Due to this, and if the perforated interval is long, the inflow will usually be unevenly distributed,
exhibiting low velocities at the bottom and high velocities at the top (even to the extreme of flow back to the
formation). When this occurs, the gas velocities at the lower part of the wellbore might not enough to lift the
liquid upwards. Figure 9‐8 shows an example of such a case. In the configuration shown, a liquid column is
established at the well bottom. Part of the liquid flows back into the formation and gas and liquid enter through
mainly through the top part of the interval.

FIGURE 9‐8. SKETCH SHOWING A CONFIGURATION AT WELL BOTTOM WITH DISTRIBUTED INFLOW WITH DOWNHOLE LIQUID
SEPARATION AND BACK‐SEEPAGE FROM WELLBORE TO THE FORMATION

The configuration shown in Figure 9‐8 can still be modeled with a steady‐state approximation.

201
 Gas well liquid loading M. Stanko

Some transient situations

There could be some situations for which a steady‐state approximation does not work. For example, situations
where part of the liquid does not flow upwards with the gas and accumulates at the bottom of the well (e.g.
in a stagnant column). However, for this situation to occur, the gas velocity must usually be very low (which
imply very low gas production rates) and the well gets filled with liquid very quickly. In these cases, one should
notice a significant drop in the liquid production of the well (e.g. a reduction in the producing WGR, CGR or
LGR).
Churn and slug flow can create transient rapid pressure fluctuations that can cause liquid flowback to the
formation and transient conditions downhole, especially if the inflow is distributed over a distance.
There could be some cases where liquid loaded wells exhibit transient production in large time scales (hours).
Some examples are listed next:

 If the well is deviated, has a horizontal section, or has a high degree of tortuosity in the horizontal
section, a phenomenon similar to severe slugging in risers and pipeline is triggered due to the
accumulation of liquids in low points. A cyclic pressure build‐up and blow down occurs that drives
liquids upwards cyclically.
 Cyclic wellbore‐formation interaction, due to, for example liquid accumulation in the near wellbore
region (condensation banking).

202
Gas well liquid loading M. Stanko

REFERENCES
[9‐1] Barnea, D. (1986), "Transition from annular flow and from dispersed bubble flow‐unified models for
the whole range of pipe inclinations", International journal of multiphase flow 12(5), 733‐744.
[9‐2] Turner, R., Hubbard, M. and Dukler, A. (1969), "Analysis and prediction of minimum flow rate for the
continuous removal of liquids from gas wells", Journal of Petroleum Technology 21(11), 1‐475.
[9‐3] Van't Westende, J. M. C. (2008), “Droplets in annular‐dispersed gas‐liquid pipe‐flows”, PhD thesis,
Delft University of Technology, Netherlands.

203
Appendices M. Stanko

APPENDICES

204
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

A. THE TUBING RATE EQUATION IN VERTICAL AND DEVIATED GAS‐WELLS

Author: Prof. Michael Golan

DERIVATION FROM FIRST PRINCIPLES (PURE SI SYSTEM)

Neglecting the acceleration term in the momentum equation, the pressure gradient at any point in the pipe is
the sum of the hydrostatic and the frictional gradients:

𝑑𝑝 𝜌∙𝑢 EQ. A‐1

𝜌 ∙ 𝑔 ∙ cos 𝛼 𝑓∙
𝑑𝑙 2∙𝐷
Where:
𝑓 It could be Moody and Fannin friction factor (fM or fF)27
𝛼 Inclination angle from the vertical direction

When the units are in British Engineering unit system, the equation becomes:

𝑑𝑝 𝑔 𝜌∙𝑢
𝜌∙ ∙ cos 𝛼 𝑓 ∙
𝑑𝑙 𝑔 2∙𝑔 ∙𝐷
and in oil field unit system, where pressure is expressed in psia, it is written as

𝑑𝑝 𝑔 𝜌∙𝑢
144 ∙ 𝜌∙ ∙ cos 𝛼 𝑓 ∙
𝑑𝑙 𝑔 2∙𝑔 ∙𝐷
Returning to the SI equation, expressing the density in terms of the compressibility factor, and the flow velocity
in terms of mass flow rate,𝑢 , gives:
∙

𝑑𝑝 𝑝∙𝑀 8∙𝑓 ∙𝑚 𝑍∙𝑅∙𝑇 EQ. A‐2

∙ 𝑔 ∙ cos 𝛼 ∙
𝑑𝑙 𝑍∙𝑅∙𝑇 𝜋 ∙𝐷 𝑝∙𝑀

Defining:

𝑀 EQ. A‐3
𝐶 ∙ 𝑔 ∙ cos 𝛼
𝑍∙𝑅∙𝑇
And

8∙𝑓 ∙𝑚 𝑍∙𝑅∙𝑇 EQ. A‐4

𝐶 ∙
𝜋 ∙𝐷 𝑀

and substituting Eq. A‐3 and Eq. A‐4 into Eq. A‐2 gives

𝐶 EQ. A‐5
𝑑𝑝 𝐶 ∙𝑝 ∙ 𝑑𝑙
𝑝
Or

𝑑𝑝 𝑝 ∙ 𝑑𝑝 EQ. A‐6
𝑑𝑙
𝐶 𝐶 ∙𝑝 𝐶
𝐶 ∙𝑝
𝑝

∙ ∙
27
The Fanning friction factor is defined as 𝑓 and the Moody friction factor 𝑓 , thus 𝑓 4∙𝑓
∙ ∙

205
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

To integrate this equation a new variable U, is defined,

𝑈 𝐶 ∙𝑝 𝐶 EQ. A‐7

𝑑𝑈 2 ∙ 𝐶 ∙ 𝑝 ∙ 𝑑𝑝 EQ. A‐8

The U and dU substituted into Eq. A‐6 give:

1 𝑑𝑈 EQ. A‐9
𝑑𝑙 ∙
2∙𝐶 𝑈
Integrating Eq. A‐9 between two points in the pipe 1 and 2, assuming that parameters Ca and Cb are constant28
between 1‐2:

1 𝑑𝑈 EQ. A‐10
𝑑𝑙 ∙
2∙𝐶 𝑈

This integral gives

1 𝑈 1 𝐶 ∙𝑝 𝐶 EQ. A‐11
𝑙 𝑙 𝐿 ∙ ln ∙ ln
2∙𝐶 𝑈 2∙𝐶 𝐶 ∙𝑝 𝐶

Or:

𝐶 ∙𝑝 𝐶 EQ. A‐12
∙ ∙
𝑒
𝐶 ∙𝑝 𝐶

Defining:

𝑀 EQ. A‐13
𝑆 2∙𝐿∙𝐶 2∙ ∙ 𝐿 ∙ 𝑔 ∙ cos 𝛼
𝑍 ∙𝑅∙𝑇
Eq. A‐12 becomes:

𝐶 ∙𝑝 𝐶 EQ. A‐14
𝑒
𝐶 ∙𝑝 𝐶

Which can be rearranged such that:

𝐶 EQ. A‐15
𝑝 𝑝 ∙𝑒 ∙ 𝑒 1
𝐶
Dividing Eq. A‐4 by Eq. A‐3 gives:
EQ. A‐16
𝐶 8∙𝑓 , ∙𝑚 𝑍 ∙𝑅∙𝑇
∙
𝐶 𝜋 ∙ 𝐷 ∙ 𝑔 ∙ cos 𝛼 𝑀

Converting the mass flow rate to volumetric flow‐rate expressed at standard conditions using Eq. A‐17

𝑝 𝑀 EQ. A‐17
𝑚 𝜌 ∙𝑞 ∙ ∙𝑞
𝑇 𝑅

28
Evaluated with average deviation factor Zav = 0.5·(Z1 + Z2), average temperature and average friction factor

206
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

results in:

𝐶 8∙𝑓 , ∙ 𝑍 ∙𝑇 𝑝 EQ. A‐18

∙ ∙𝑞
𝐶 𝜋 ∙ 𝐷 ∙ 𝑔 ∙ cos 𝛼 𝑇
Substituting Eq. A‐18 into Eq. A‐15:

8∙𝑓 𝑝 EQ. A‐19

𝑝 𝑝 ∙𝑒 ∙ ∙ 𝑍 ∙𝑇 ∙ 𝑒 1 ∙𝑞
𝜋 ∙ 𝐷 ∙ 𝑔 ∙ cos 𝛼 𝑇
Multiplying and dividing the second term on the right‐hand side with:

𝑀 ∙𝑔 28.97 ∙ 𝛾 ∙ 𝑔 EQ. A‐20

𝑆 2∙ ∙ 𝐿 ∙ cos 𝛼 2∙ ∙ 𝐿 ∙ cos 𝛼
𝑍 ∙𝑅∙𝑇 𝑍 ∙𝑅∙𝑇
28.97 𝑝 𝑒 1 EQ. A‐21
𝑝 𝑝 ∙𝑒 ∙ 𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 ∙ ∙𝑞
𝑅 𝑇 𝑆∙𝐷
Solving for the flow rate:

. . . EQ. A‐22
𝜋 𝑅 𝑇 𝐷 𝑆
𝑞 ∙ ∙ ∙ ∙ 𝑝 𝑝 ∙𝑒 ∙
4 𝑀 𝑝 𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 𝑒 1

This equation relates the pressure at the top and the bottom of the tubing.
In fully turbulent flow (high Reynolds numbers), the friction factor depends essentially on the relative
roughness of the pipe, ε/D, and becomes independent of the Reynolds number. Measurements in gas wells
conducted by R.V.Smith, (1950), yielded a correlation for friction factor in tubings that became the norm for
most equations used by the gas industry and which appear in engineering handbooks. Smith’s measurements
are expressed in terms of friction factor as:

0.01748 0.0077 EQ. A‐23

𝑓 .
. 39.37 𝑖𝑛 𝐷 .
𝐷 ∙ |1 𝑚| ∙
1𝑚
By defining the tubing constant 𝐶 :

. . . EQ. A‐24
𝜋 𝑅 𝑇 𝐷 𝑆∙𝑒
𝐶 ∙ ∙ ∙ ∙
4 𝑀 𝑝 𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 𝑒 1

This yields:
. EQ. A‐25
𝑝
𝑞 𝐶 ∙ 𝑝 ∙
𝑒

PRESSURE EQUATION IN PRACTICAL FIELD UNITS (METRIC)

Eq. A‐22 is rearranged to make it explicit with respect to bottom‐hole pressure:

16 28.97 𝑝 𝑒 1 EQ. A‐26

𝑝 𝑝 ∙𝑒 ∙ ∙ 𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 ∙ ∙𝑞
𝜋 𝑅 𝑇 𝑆∙𝐷
Or:

207
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

𝑒 1 EQ. A‐27
𝑝 𝑝 ∙𝑒 658 ∙ 𝑓 ∙ 𝐿 ∙ 𝛾 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
𝑆∙𝐷
When converting to practical metric units, Sm3/d, bara, m, the equation becomes:

𝑒 1 EQ. A‐28
𝑝 𝑝 ∙𝑒 8.8 ∙ 10 ∙𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 ∙ ∙𝑞
𝑆∙𝐷
Alternatively, if one chooses to introduce the definition of S (from Eq. A‐20) in Eq. A‐22, and rearrange it to
make it explicit with respect to bottom‐hole pressure:

8∙𝑓 𝑝 EQ. A‐29

𝑝 𝑝 ∙𝑒 ∙ ∙ 𝑍 ∙𝑇 ∙ 𝑒 1 ∙𝑞
𝜋 ∙ 𝐷 ∙ 𝑔 ∙ cos 𝛼 𝑇
when substituting values for the constants:
EQ. A‐30
8 10 𝑒 1
𝑝 𝑝 ∙𝑒 ∙ ∙𝑓 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
9.81 ∙ 𝜋 293 𝐷 ∙ cos 𝛼

Giving

𝑒 1 EQ. A‐31
𝑝 𝑝 ∙𝑒 9624 ∙ 𝑓 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
𝐷 ∙ cos 𝛼
When converting to practical metric units, Sm3/d, bara, m, the equation becomes

9624 𝑒 1 EQ. A‐32

𝑝 𝑝 ∙𝑒 ∙𝑓 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
10 ∙ 86400 𝐷 ∙ cos 𝛼
Or

𝑒 1 EQ. A‐33
𝑝 𝑝 ∙𝑒 1.295 ∙ 10 ∙𝑓 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
𝐷 ∙ cos 𝛼

The tubing gas equation expressed in terms of the bottom‐hole pressure translated to wellhead datum level
(Fetkovich approach)

In integrated gas field studies, it is convenient to analyze the flow of the entire production system using the
wellhead or the top of the well as a reference datum level. Mike Fetkovich has suggested this approach in a
1975 paper. He rearranged Eq. A‐22 as follows:

. . . EQ. A‐34
𝜋 𝑅 𝑇 𝐷 𝑝 𝑆∙𝑒
𝑞 ∙ ∙ ∙ ∙ 𝑝 ∙
4 𝑀 𝑝 𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 𝑒 𝑒 1

Substituting 𝑝

, . . EQ. A‐35
𝜋 𝑅 𝑇 𝐷 𝑆∙𝑒
𝑞 ∙ ∙ ∙ ∙ 𝑝 𝑝 ∙
4 𝑀 𝑝 𝑓 ∙𝐿∙𝛾 ∙𝑍 ∙𝑇 𝑒 1

Where pw represents the flowing bottom hole pressure expressed at wellhead datum level. The quantity, pw
is actually the bottom‐hole flowing pressure minus the hydrostatic pressure of the gas column.
Substituting the definition of S (Eq. A‐20) (all in pure SI system):

208
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

. / EQ. A‐36
𝜋 𝑇 ,
𝐷 𝑒 .
𝑞 ∙ ∙ 2 ∙ 𝑔 ∙ cos 𝛼 ∙ ∙ ∙ 𝑝 𝑝
4 𝑝 𝑓 𝑍 ∙𝑇 ∙ √𝑒 1
In practical metric units, where: qsc in [Sm3/d], pressure in [bara], length in [m] and temperature in [K], the
equation becomes:
. / EQ. A‐37
86400 ∙ 𝜋 .
288 𝐷 𝑒
𝑞 ∙ 2 ∙ 9.81 ∙ cos 𝛼 ∙ ∙ ∙
4 1 𝑓 𝑍 ∙𝑇 ∙ √𝑒 1
∙ 𝑝 𝑝 .
Or
. / EQ. A‐38
.
𝐷 𝑒 .
𝑞 86.56 ∙ 10 ∙ cos 𝛼 ∙ ∙ ∙ 𝑝 𝑝
𝑓 𝑍 ∙𝑇 ∙ √𝑒 1
In vertical wells H = L, and cos(α) = 1. When substituting into the rate equation, together with the expression
for fully turbulent friction factor (Eq. A‐23) gives:

𝐷 .
∙𝑒 / EQ. A‐39
.
𝑞 0.986 ∙ 10 ∙ ∙ 𝑝 𝑝
𝑍 ∙𝑇 ∙ √𝑒 1
This is the metric version of the rate equation suggested by Fetkovich for integrated field studies.
In oilfield units (psia, MSCFD, ft, R), the datum corrected rate equation (Eq. A‐36) is:
. / EQ. A‐40
𝜋 520 .
𝐷 𝑒
𝑞 86.4 ∙ ∙ ∙ 2 ∙ 32.17 ∙ cos 𝛼 ∙ ∙
4 14.7 12 ∙ 𝑓 𝑍 ∙𝑇 ∙ √𝑒 1
∙ 𝑝 𝑝 .

Substituting the expression for fully turbulent Moody friction factor (Eq. A‐23):
, EQ. A‐41
𝜋 520 ,
𝐷 . 𝑒 /
𝑞 693.034 ∙ ∙ ∙ cos 𝛼 ∙ ∙
4 14.7 12 ∙ 0.01748 𝑍 ∙𝑇 ∙ √𝑒 1
∙ 𝑝 𝑝 .
Which finally gives:

𝐷 .
∙𝑒 / EQ. A‐42
.
𝑞 292.9 ∙ ∙ 𝑝 𝑝
𝑍 ∙𝑇 ∙ √𝑒 1

FETKOVICH RATE EQUATION

The equation used by Fetkovich in his 1975 is derived from the IOCC manual and is (rate is in MSCFD):

31.62 ∙ 𝑒 / EQ. A‐43

.
𝑞 ∙ 𝑝 𝑝
𝐹 ∙𝑍 ∙𝑇 ∙ √𝑒 1
.
where 𝐹 .

209
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

The relationship between Fr29and the friction factors is, by definition:

𝑓 EQ. A‐44
2.6665 ∙ 𝑓 ∙ 𝑞 2.6665 ∙ ∙𝑞
𝐹 4
𝐷 𝐷
Where D is inner tubing diameter, in and q is the gas rate in MMSCFD.
By substituting the empirical value of Fr to the rate equation it becomes:

𝐷 .
∙𝑒 / EQ. A‐45
.
𝑞 292.9 ∙ ∙ 𝑝 𝑝
𝑍 ∙𝑇 ∙ √𝑒 1
For verification purposes, the equation will be converted to practical metric units

1000 1 39.37 ∙ 𝐷 .
∙𝑒 /
14.7 EQ. A‐46
.
𝑞 292.9 ∙ ∙ ∙ ∙ 𝑝 𝑝
1 35.14 𝑍 ∙ 1.8 ∙ 𝑇 ∙ √𝑒 1 1
Or:

𝐷 .
∙𝑒 / EQ. A‐47
.
𝑞 0.986 ∙ 10 ∙ ∙ 𝑝 𝑝
𝑍 ∙𝑇 ∙ √𝑒 1
The relationship between fM and the Fr in the IOCC equation:
Interstate Oil Compact Commission “manual of Backpressure Testing of Gas Wells”, Oklahoma City, Oklahoma
Cullender and Smith (1956) introduced originally the dimensional expression Fr. It is a function of fM, flow rate,
and pipe diameter. Back calculating the friction factor from the Fr used in the IOCC equation yields

0.00437 EQ. A‐48

𝑓
𝐷 .
0.01748 EQ. A‐49
𝑓
𝐷 .
Starting with the IOCC equation as listed in Fetkovich’s paper from 1975 (before dividing by eS for datum
change):

𝑞∙𝐹 ∙𝑇 ∙𝑍 EQ. A‐50

𝑝 𝑝 ∙𝑒 ∙ 𝑒 1
31.62
Rearranging:

𝐹 ∙𝑇 ∙𝑍 EQ. A‐51
𝑝 𝑝 ∙𝑒 ∙ 𝑒 1 ∙𝑞
31.62
where:
.
𝐹 . , pipe diameter D in inch, and the gas rate in MSCFD.

(Note: there is an error in the pressure equation in the original 1975 paper where the equations are hand
written, there the number 31.62 is wrongly written as 1.000. The error has been corrected in later prints of

29
The dimensional expression Fr has been introduced originally by Cullender and Smith (1956) to facilitate another
method to calculated bottom hole pressure accounting for changes in temperature and compressibility factor. The IOCC
preferred to apply it in its manual rather than the dimensionless friction factor (Oklahoma City People versus the rest of
the world).

210
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

the paper, also be aware that the rate equation in most gas engineering manuals is reported in MMSCFD,
Fetkovich uses MSCFD in his analysis)
For comparison, taking any of the widely used engineering equations, for example in the SPE –Petroleum
Engineering Handbook (Chapter 34 “Wellbore Hydraulics” by Bertuzzi, Fetkovich, Poettmann and Thomas,
equation 44) which applies Moody friction factor fm:

𝑒 1 EQ. A‐52
𝑝 𝑝 ∙𝑒 25 ∙ 𝑓 ∙ 𝐻 ∙ 𝛾 ∙ 𝑍 ∙ 𝑇 ∙ ∙𝑞
𝑆∙𝐷
or, by substituting the expression for S:

𝑒 1 EQ. A‐53
𝑝 𝑝 ∙𝑒 25 ∙ 𝑓 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
0.0375 ∙ 𝐷
or in the The Canadian Energy Resource Conservation Board Manual (ERCB) on gas well testing which applies
Fanning friction factor (note that Moody factor is 4 times Fanning factor):

𝑒 1 EQ. A‐54
𝑝 𝑝 ∙𝑒 100 ∙ 𝑓 ∙ 𝐻 ∙ 𝛾 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
𝑆∙𝐷
The units in these two equations are: P in [psia], vertical depth H in [ft], q = flow‐rate in [MMSCFD], d in [in],
friction factor f [‐], and S is expressed by the following expression:

28.97 ∙ 𝛾 ∙ 𝑔 28.97 ∙ 32.174 𝛾 EQ. A‐55

𝑆 2∙ ∙𝐻 2 ∙ ∙𝐻
𝑍 ∙𝑇 ∙𝑅 10.732 ∙ 144 ∙ 32.174 𝑍 ∙ 𝑇
𝛾
0.0375 ∙ ∙𝐻
𝑍 ∙𝑇
To back‐calculate the friction factor as implied by the IOCC equation, a comparison is made between the
second terms on the right‐hand side of the IOCC and the ERCB equations (converting it from MMSCFD to
MSCFD as used by the IOCC).

𝐹 ∙𝑍 ∙𝑇 𝑒 1 EQ. A‐56
∙ 𝑒 1 ∙𝑞 100 ∙ 𝑓 ∙ 𝐻 ∙ 𝛾 ∙ 𝑍 ∙𝑇 ∙ ∙ 𝑞 ∙ 10
31.62 𝑆∙𝐷
Substituting s in the denominator of the right‐hand side gives:

1000 ∙ 𝐹 ∙ 𝑍 ∙𝑇 ∙ 𝑒 1 ∙𝑞 EQ. A‐57

𝑒1
100 ∙ 𝑓 ∙ 𝐻 ∙ 𝛾 ∙ 𝑍 ∙𝑇 ∙ ∙𝑞
𝛾 ∙𝐻
0.0375 ∙ ∙𝐷
𝑍 ∙𝑇
which, when compared with the relevant term in IOCC equation gives:

1 EQ. A‐58
𝐹 0.1 ∙ 𝑓 ∙
0.0375 ∙ 𝐷
Solving for the Fanning Friction factor, 𝑓

𝑓 0.375 ∙ 𝐷 ∙ 𝐹 EQ. A‐59

.
and substituting 𝐹 . gives:

211
 Appendix A: Tubing Rate Equation in Vertical and Deviated Gas Wells M. Golan

0.10797 0.00437 EQ. A‐60

𝑓 0.375 ∙ 𝐷 ∙
𝐷 . 𝐷 .

which is equivalent to:

0.10797 0.0174 EQ. A‐61

𝑓 4∙ ∙ 0.375 ∙ 𝐷
𝐷 . 𝐷 .
The diameter, 𝐷, in both expressions is in inch (While the pipe length in the equation is in ft).

REFERENCES

 Katz, D.L., Cornel, D., Kobayashi, R., Poetmann, F.H., Vary, J.A. Elenbass, J.R., Weinaug, C.F. (1959).
Handbook of Natural Gas Engineering. McGraw‐Hill Publishing Company.
 Smith, R.V. (1950). Determining Friction Factors for Measuring Productivity of Gas Wells. Trans AIME
189 (73).
 Energy Resources Conservation Board (ERCB) (1975). Theory and Practice of the Testing of Gas Wells.
ERCB 73‐34, Third Edition, 1975.
 Katz, D.L., & Lee, R.L., (1990). Natural Gas Engineering‐Production and Storage. McGraw‐Hill
Publishing Company.
 Young, K.L. (1966). Effect of Assumptions Used to Calculate Bottom‐Hole Pressures in Gas Wells. SPE‐
1626. Society of Petroleum Engineers
 Smith, R.V., Williams R.H., & Dewees, E.J. (1954). Measurements of Resistance to Flow of Fluids in
Natural Gas Wells. Trans AIME 201, pp. 279
 Cullender M.H. & Smith R.V. (1956) Practical solution of the Gas Flow Equations for Wells and Pipelines
with Large Temperature Gradients. Trans AIME 207 pp. 281‐287.

212
 Appendix B: Choke Equations M. Stanko

B. CHOKE EQUATIONS

UNDERSATURATED OIL FLOW

Based on a frictionless flow contraction from an upstream point 1 to a downstream point 2.

The single‐phase Bernoulli equation for steady state frictionless flow along a streamline, neglecting elevation
changes, is:

𝑑𝑝 EQ. B‐1
𝑉 ∙ 𝑑𝑉 0
𝜌
Where:
𝑝 Pressure
𝜌 Density
𝑉 Velocity

Integrating Eq. B‐1 from point 1 to 2:

𝑑𝑝 EQ. B‐2
𝑉 ∙ 𝑑𝑉 0
𝜌

Assuming incompressible flow:

𝑝 𝑝 𝑉 𝑉 EQ. B‐3
0
𝜌 2
The mass is conserved in the choke, thus:

𝑉 ∙𝐴 𝑉 ∙𝐴 EQ. B‐4

The area upstream the choke can be expressed with the diameter of the pipe upstream the choke:

𝜋∙∅ EQ. B‐5

𝐴
4
In a similar way, the cross‐section area of 2:

𝜋∙∅ EQ. B‐6

𝐴
4
Using Eq. B‐4, Eq. B‐5 and Eq. B‐6, it is possible to express V1 as a function of V2:

𝐴 ∅ EQ. B‐7
𝑉 𝑉 ∙ 𝑉 ∙
𝐴 ∅

To simplify the nomenclature, the ratio between the diameters is named beta (which, in a contraction, is
always less than 1):

∅ EQ. B‐8
𝛽
∅
Substituting Eq. B‐7 in Eq. B‐3:

213
 Appendix B: Choke Equations M. Stanko

𝑝 𝑝 𝑉 𝑉 ∙𝛽 EQ. B‐9
𝜌 2
Clearing V2 in Eq. B‐9:
EQ. B‐10
2∙ 𝑝 𝑝
𝑉
𝜌∙ 1 𝛽

For petroleum production calculations, we often require the oil rate at standard conditions, not the velocity,
thus, multiplying Eq. B‐10 by A2 and the oil volume factor Bo,@2:
EQ. B‐11
𝐴 2∙ 𝑝 𝑝
𝑞 ∙
𝐵 ,@ 𝜌∙ 1 𝛽

Where Bo,@2 and ρ are evaluated at p2 and T2.

As mentioned earlier, due to the “vena contracta” effect, the effective area at the throat is not exactly A2, but
slightly less. Thus, a correction factor called the flow coefficient is introduced in Eq. B‐11:
EQ. B‐12
𝐴 ∙𝐶 2∙ 𝑝 𝑝
𝑞 ∙
𝐵 ,@ 𝜌∙ 1 𝛽

DRY GAS FLOW

(based on a frictionless flow contraction from an upstream point 1 to a downstream point 2)
Using Eq. B‐2 as the starting point, the term related to pressure and density remains valid; however, in gas
flow the velocity downstream is usually much higher than the velocity upstream, thus V22 >> V12:

𝑑𝑝 𝑉 EQ. B‐13
0
𝜌 2

The density will vary inside the choke. An assumption commonly used is that the contraction process is
adiabatic (with an exponent k, the ratio between the specific heats of the gas):

𝑝∙𝜌 𝐶 EQ. B‐14

Where C is a constant. Substituting Eq. B‐14 in Eq. B‐13:

𝑑𝑝 𝑉 EQ. B‐15
𝐶 ∙ 0
2
𝑝
Solving the integral:

𝑘 𝑉 EQ. B‐16
𝐶 ∙ ∙ 𝑝 𝑝 0
𝑘 1 2

The constant C is expressed in terms of the inlet conditions:

EQ. B‐17
𝑝
𝐶
𝜌

214
 Appendix B: Choke Equations M. Stanko

Substituting Eq. B‐17 in Eq. B‐16 and introducing the pressure ratio y = p2/p1:
EQ. B‐18
𝑝 𝑘 𝑉
∙ ∙𝑝 ∙ 𝑦 1 0
𝜌 𝑘 1 2
Clearing V2 and simplifying p1:
EQ. B‐19
𝑝 𝑘
𝑉 2∙ ∙ ∙ 1 𝑦
𝜌 𝑘 1

Expressing ρ1 with the real gas equation:

𝑝 ∙𝑀 EQ. B‐20
𝜌
𝑍 ∙𝑅∙𝑇
Where:
𝑀 Molecular weight of the gas
𝑅 Universal gas constant
𝑍 Generalized compressibility factor

Substituting Eq. B‐20 in Eq. B‐19:

EQ. B‐21
𝑍 ∙𝑅∙𝑇 𝑘
𝑉 2∙ ∙ ∙ 1 𝑦
𝑀 𝑘 1

For petroleum production calculations, we often require the gas rate at standard conditions, not the velocity,
thus, multiplying Eq. B‐21 by the “effective” cross‐section area of 2 gives the local volume rate:
EQ. B‐22
𝑍 ∙𝑅∙𝑇 𝑘
𝑞 𝐴 ∙𝐶 ∙ 2∙ ∙ ∙ 1 𝑦
𝑀 𝑘 1

The local volumetric rate at point 2 is related to the rate at standard conditions by the following equation:

𝑞 ∙𝜌 𝑞 ∙𝜌 EQ. B‐23

Substituting Eq. B‐23 in Eq. B‐22 gives:

EQ. B‐24
𝜌 ∙𝐴 ∙𝐶 𝑍 ∙𝑅∙𝑇 𝑘
𝑞 ∙ 2∙ ∙ ∙ 1 𝑦
𝜌 𝑀 𝑘 1

ρ2 is related with ρ1 by Eq. B‐17:

EQ. B‐25
𝑝 𝑝
𝜌 𝜌
Clearing ρ2 from Eq. B‐25 and substituting in Eq. B‐24, and using the real gas equation to express the gas
density at standard conditions:

215
 Appendix B: Choke Equations M. Stanko

EQ. B‐26
𝜌 ∙𝑝 ∙𝑅∙𝑇 ∙𝐴 ∙𝐶 𝑍 ∙𝑅∙𝑇 𝑘
𝑞 ∙ 2∙ ∙ ∙ 1 𝑦
𝑀 𝑘 1
𝑝 ∙𝑝 ∙𝑀

Introducing Eq. B‐20 for ρ1:

EQ. B‐27
𝑝 ∙𝑀 ∙𝑝 ∙𝑅∙𝑇 ∙𝐴 ∙𝐶 𝑍 ∙𝑅∙𝑇 𝑘
𝑞 ∙ 2∙ ∙ ∙ 1 𝑦
𝑀 𝑘 1
𝑍 ∙𝑅∙𝑇 ∙𝑝 ∙𝑝 ∙𝑀

Simplifying and rearranging terms:

EQ. B‐28
𝑝 ∙𝑇 ∙𝐴 ∙𝐶 𝑅 𝑘
𝑞 ∙ 2∙ ∙ ∙ 𝑦 𝑦
𝑝 𝑍 ∙𝑇 ∙𝑀 𝑘 1

Cd depends on the geometry of the restriction, the Reynolds number and the ratio between the upstream and
downstream diameters. If no information is available, a value of 0.865 can be used.
Eq. B‐28 is valid only for the subcritical range. To predict the rate in the critical range the critical pressure ratio
(yc) must be used, instead of the actual pressure ratio.
For gas, the critical pressure ratio can be predicted as:
EQ. B‐29
2
𝑦
𝑘 1
OIL‐GAS‐WATER MIXTURE
There is often a mixture of oil, gas and water circulating through the choke. To estimate fluid properties at the
choke outlet or at the throat, an assumption that is typically made is that the mixture undergoes an adiabatic
expansion. Using the first law of the thermodynamics and assuming piston work yields:

𝑑𝑢 𝑝 ∙ 𝑑𝑣 EQ. B‐30

Where:
𝑢 Specific internal energy
𝑝 pressure
𝑣 Specific volume of the mixture

The variation in specific internal energy is expressed in terms of the specific heat at constant volume:

𝑑𝑢 𝑥 ∙𝐶 𝑥 ∙𝐶 𝑥 ∙𝐶 ∙ 𝑑𝑇 EQ. B‐31
, , ,

Where:
𝑥 Molar fraction of phase “𝑖”
𝐶 , Specific heat at constant volume of phase “𝑖”
𝑇 temperature

Or, introducing the specific heat at constant volume of the mixture:

216
 Appendix B: Choke Equations M. Stanko

𝑑𝑢 𝐶 , ∙ 𝑑𝑇 EQ. B‐32

The specific volume of the mixture is expressed in terms of the mixture density:
𝑝 EQ. B‐33
𝑑𝑤 𝑝 ∙ 𝑑𝑣 ∙ 𝑑𝜌
𝜌
Substituting Eq. B‐32 and Eq. B‐33 in equation Eq. B‐30 yields:
𝑝 EQ. B‐34
𝐶 , ∙ 𝑑𝑇 ∙ 𝑑𝜌
𝜌
We wish to express temperature as a function of pressure and mixture density. The density of the mixture is:

1 EQ. B‐35
𝜌 𝑥 𝑥 𝑥
𝜌 𝜌 𝜌

Clearing the gas density:

1 1 1 𝑥 𝑥 EQ. B‐36
∙
𝜌 𝑥 𝜌 𝜌 𝜌

Substituting the ideal gas equation:

𝑝 1 𝑥 𝑥 EQ. B‐37
𝑇 ∙
𝑅∙𝑥 𝜌 𝜌 𝜌

Deriving the expression (assuming that oil and water densities and molar fractions remain constant during the
choke expansion):

𝑑𝑝 1 𝑥 𝑥 𝑝 1 EQ. B‐38
𝑑𝑇 ∙ ∙ ∙ 𝑑𝜌
𝑅∙𝑥 𝜌 𝜌 𝜌 𝑅∙𝑥 𝜌

Substituting in Eq. B‐34 yields:

𝑑𝑝 1 𝑥 𝑥 𝑝 1 𝑝 EQ. B‐39
𝐶 , ∙ ∙ ∙ ∙ 𝑑𝜌 ∙ 𝑑𝜌
𝑅∙𝑥 𝜌 𝜌 𝜌 𝑅∙𝑥 𝜌 𝜌

Rearranging terms (all that is related to pressure to the right side and with density to the left):

𝑅∙𝑥 𝑑𝜌 𝑑𝑝 EQ. B‐40

𝜌 ∙𝜌 ∙ 1 ∙ ∙
𝐶, 𝜌 ∙𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌 𝑝

Performing an indefinite integration on both sides and subsequently taking the exponential:
∙ , EQ. B‐41
𝜌 ,
𝑝∙ 𝑐∙
𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌
Where 𝑐 is a constant.
Substituting the definition of specific heat at constant volume of the mixture (Eq. B‐32):
∙ ∙ , ∙ , ∙ , EQ. B‐42
𝜌 ∙ , ∙ , ∙ ,
𝑝∙ 𝑐∙
𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌

217
 Appendix B: Choke Equations M. Stanko

And using the relationship between the gas specific heat at constant volume, specific heat at constant pressure
and the universal gas constant:

𝑅 𝐶 𝐶 EQ. B‐43

Yields:
∙ , ∙ , ∙ , EQ. B‐44
𝜌 ∙ , ∙ , ∙ ,
𝑝∙ 𝑐
𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌
Eq. B‐36 is rearranged to express the density of the gas as a function of the density of the mixture:
𝜌 𝜌 EQ. B‐45
𝑐
𝑥 ∙𝜌 ∙𝜌 𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌 𝑥 ∙𝜌 ∙𝜌

Substituting Eq. B‐45 in Eq. B‐44 yields:

∙ , ∙ , ∙ , EQ. B‐46
𝜌 ∙ , ∙ , ∙ ,
𝑝∙ 𝑐
𝑥 ∙𝜌 ∙𝜌

The gas mole fraction, oil density and water density are assumed to remain constant during the choke
expansion, therefore:
∙ , ∙ , ∙ , EQ. B‐47
∙ , ∙ ∙ ,
𝑝∙ 𝑣 , 𝑐

This equation resembles a polytropic process across the choke.

218
 Appendix D: Temperature Drop in Conduit M. Stanko

C. TEMPERATURE DROP IN CONDUIT FOR LIQUID FLOW

GENERAL EXPRESSION

Departing from the general steady state energy equation in one dimension:

𝑑𝑞 𝑑ℎ 𝑑𝑣 𝑑𝑧 EQ. C‐1
𝑚∙ 𝑣∙ 𝑔∙
𝑑𝐿 𝑑𝐿 𝑑𝐿 𝑑𝐿
The heat transfer can be expressed with the overall heat transfer coefficient, based on the outer diameter of
the conduit:

𝑞 2∙𝜋∙𝐿∙𝑟 ∙𝑈 ∙ 𝑇 𝑇 EQ. C‐2

,

Where:
𝑟 , Insulation outer radius [m]
𝑈 Overall heat transfer coefficient [W/m2.K]
𝑇 Mean ambient temperature [K or °C]
𝑇 Mean fluid temperature in the section [K or °C]

Differentiating Eq. C‐2:

𝑑𝑞 2∙𝜋∙𝑟 ∙𝑈 ∙ 𝑇 𝑇 ∙ 𝑑𝐿 EQ. C‐3

,

Making Tf = T and neglecting velocity changes 0, Eq. C‐1 becomes:

𝑑ℎ 𝑑𝑧 EQ. C‐4
2∙𝜋∙𝑟 , ∙𝑈 ∙ 𝑇 𝑇 𝑚∙ 𝑔∙
𝑑𝐿 𝑑𝐿
DERIVATION FOR LIQUIDS
For liquids, assuming incompressibility, enthalpy can be expressed in terms of the specific heat capacity at
constant pressure:

𝑑ℎ 𝐶 ∙ 𝑑𝑇 EQ. C‐5

Where:
𝐶 Specific heat capacity [J/kg.K]

Heat entering to the system is positive, heat leaving negative, thus a negative sign must be placed in front of
the heat expression.

𝑑𝑇 EQ. C‐6
2∙𝜋∙𝑟 , ∙𝑈 ∙ 𝑇 𝑇 𝑚∙ 𝐶 ∙ 𝑔 ∙ sin 𝜃
𝑑𝐿
Where:
𝜃 Angle between pipe and horizontal [rad]

Expanding the expression:

219
 Appendix D: Temperature Drop in Conduit M. Stanko

𝑑𝑇 EQ. C‐7
𝑇∙2∙𝜋∙𝑟 , ∙𝑈 𝑇 ∙2∙𝜋∙𝑟 , ∙𝑈 𝑚∙𝐶 ∙ 𝑚 ∙ 𝑔 ∙ sin 𝜃
𝑑𝐿
𝑑𝑇 2𝜋 ∙ 𝑟 , ∙ 𝑈 𝑇 ∙2∙𝜋∙𝑟 , ∙𝑈 𝑔 ∙ sin 𝜃 EQ. C‐8
𝑇∙ 0
𝑑𝐿 𝑚∙𝐶 𝑚∙𝐶 𝐶

For simplicity, we define the variable A:

𝑚∙𝐶 EQ. C‐9

𝐴
2𝜋 ∙ 𝑟 , ∙ 𝑈
𝑑𝑇 1 𝑇 𝑔 ∙ sin 𝜃 EQ. C‐10
𝑇∙ 0
𝑑𝐿 𝐴 𝐴 𝐶

Solving the differential equation, using 𝑢 𝑒 and multiplying it by the above expression:

𝑑𝑇 1 𝑇 𝑔 ∙ sin 𝜃 EQ. C‐11

𝑢∙ 𝑢∙𝑇∙ 𝑢∙
𝑑𝐿 𝐴 𝐴 𝐶

The product differentiating rule is defined as:

𝑑 𝑤 𝑥 ∙𝑣 𝑥 𝑑𝑤 𝑥 𝑑𝑣 𝑥 EQ. C‐12
.𝑣 𝑥 𝑤 𝑥 ∙
𝑑𝑥 𝑑𝑥 𝑑𝑥
Using the result from Eq. C‐12, it is possible to group Eq. C‐11 as follows:

𝑑 𝑢∙𝑇 𝑇 𝑔 ∙ sin 𝜃 EQ. C‐13

𝑢∙
𝑑𝐿 𝐴 𝐶

The resulting expression can be integrated by separating variables between the initial position “0” to a generic
position x in the pipe. This assumes that T∞, A, θ and Cp remain constant along the pipe length:

𝑇 𝑔 ∙ sin 𝜃 EQ. C‐14

𝑒 ∙ 𝑑𝑇 ∙ 𝑒 ∙ 𝑑𝑥
𝐴 𝐶
𝑇 𝑔 ∙ sin 𝜃 EQ. C‐15
𝑒 ∙𝑇 ∙𝐴∙𝑒
𝐴 𝐶
Where:
𝑇 Temperature of the fluid at pipe inlet [K or °C]

Evaluating at the integration limits:

𝑇 𝑔 ∙ sin 𝜃 EQ. C‐16

𝑒 ∙ 𝑇 𝑥 𝑇 ∙𝐴∙ 𝑒 1
𝐴 𝐶
𝑇 𝑔 ∙ sin 𝜃 EQ. C‐17
𝑇 𝑥 𝑇 ∙𝑒 ∙𝐴∙ 1 𝑒
𝐴 𝐶

This gives finally:

1 EQ. C‐18
𝑇 𝑥 𝑇 𝑇 𝑥 0 𝑇 ∙𝑒 ∙ 𝑔 ∙ sin 𝜃 ∙ 𝐴 ∙ 1 𝑒
𝐶

220
 Appendix D: Temperature Drop in Conduit M. Stanko

WITH VARIABLE AMBIENT TEMPERATURE

Ambient temperature could be variable, e.g. in a vertical tubing or casing along a formation. In this case, T∞
must be substituted by a function of x. Assuming a linear temperature gradient, T∞ can be expressed as:

𝑇 𝑥 𝑇 | sin 𝜃 ∙ 𝑥 ∙ 𝐺 EQ. C‐19

Where:
𝐺 Linear temperature gradient [K/m]
𝑇 | Temperature of the environment at the beginning of the section (negative if temperature
is reduced with depth) [K or °C]

Substituting in Eq. C‐14

𝑇 | sin 𝜃 ∙ 𝑥 ∙ 𝐺 𝑔 ∙ sin 𝜃 EQ. C‐20

𝑒 ∙ 𝑑𝑇 ∙ 𝑒 ∙ 𝑑𝑥
𝐴 𝐴 𝐶

221
 Appendix E: Derivation of Multiphase Flow Expressions M. Stanko

D. DERIVATION OF MULTIPHASE FLOW EXPRESSIONS

RELATIONSHIP BETWEEN LIQUID HOLDUP (HL), SLIP RATIO (S) AND NON‐SLIP LIQUID VOLUME FRACTION
(L)
Using the relationship between real (ug) and superficial (usg) gas velocities:

𝑢 ∙𝐴∙ 1 𝐻 𝑢 ∙𝐴 EQ. D‐1

Where:
𝐴 Pipe cross‐section

Using the relationship between real (ul) and superficial (usl) liquid velocities:

𝑢 ∙𝐴∙𝐻 𝑢 ∙𝐴 EQ. D‐2

Dividing Eq. D‐1 by Eq. D‐2 gives:

𝑢 ∙𝐴∙ 1 𝐻 𝑢 ∙𝐴 EQ. D‐3

𝑢 ∙𝐴∙𝐻 𝑢 ∙𝐴
Simplifying and introducing the definition of the slip ratio S = ug / ul:

𝑆∙ 1 𝐻 𝑢 EQ. D‐4
𝐻 𝑢
The non‐slip liquid volume fraction is defined as:
𝑢 EQ. D‐5
𝜆
𝑢 𝑢

or, alternatively,

1 EQ. D‐6
𝜆 𝑢
1
𝑢

Clearing out gives

𝑢 1 𝜆 EQ. D‐7
𝑢 𝜆
Substituting Eq. D‐7 into Eq. D‐4 gives

𝑆∙ 1 𝐻 1 𝜆 EQ. D‐8
𝐻 𝜆

RELATIONSHIP BETWEEN HOLDUP (HL), SLIP RATIO (S) AND QUALITY (x)

Using Eq. D‐4, the ratio of superficial velocities is expressed using the total mass flow rate (ṁ) and the quality
(gas mass fraction, x) and the densities of gas and liquid (ρl, ρg). The resulting expression is simplified:

222
 Appendix E: Derivation of Multiphase Flow Expressions M. Stanko

𝑚∙𝑥 EQ. D‐9

𝑆∙ 1 𝐻 𝜌 ∙𝐴 𝑥∙𝜌
𝐻 𝑚∙ 1 𝑥 𝜌 ∙ 1 𝑥
𝜌 ∙𝐴
Clearing HL from Eq. D‐9:

𝑆∙ 1 𝐻 1 𝑥∙𝜌 EQ. D‐10

𝑆∙ 1
𝐻 𝐻 𝜌 ∙ 1 𝑥
1 𝑥∙𝜌 EQ. D‐11
1
𝐻 𝑆∙𝜌 ∙ 1 𝑥
1 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 EQ. D‐12
𝐻 𝑆∙𝜌 ∙ 1 𝑥
𝑆∙𝜌 ∙ 1 𝑥 EQ. D‐13
𝐻
𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥

HOLDUP AVERAGE MIXTURE DENSITY (ρm)

The holdup average mixture density can be expressed as function of slip ratio (S), quality (gas mass fraction x)
and the densities of gas and liquid (ρl, ρg). The density of the mixture is defined:

𝜌 1 𝐻 ∙𝜌 𝐻 ∙𝜌 EQ. D‐14

Substituting the holdup from Eq. D‐13 in Eq. D‐14:

𝑆∙𝜌 ∙ 1 𝑥 𝑆∙𝜌 ∙ 1 𝑥 EQ. D‐15

𝜌 1 ∙𝜌 ∙𝜌
𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥

Simplifying:

𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 EQ. D‐16

𝜌 ∙𝜌 ∙𝜌
𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥
𝑥 𝑆∙ 1 𝑥 EQ. D‐17
𝜌
𝑥 𝑆
∙ 1 𝑥
𝜌 𝜌

EFFECTIVE MOMENTUM DENSITY

𝑚 EQ. D‐18
𝑚 ∙𝑢 𝑚 ∙𝑢
𝜌 ∙𝐴
1 𝐴 EQ. D‐19
∙ 𝑚 ∙𝑢 𝑚 ∙𝑢
𝜌 𝑚
1 𝐴 EQ. D‐20
∙ 𝑥∙𝑢 1 𝑥 ∙𝑢
𝜌 𝑚
1 𝐴 EQ. D‐21
∙𝑢 ∙ 𝑥∙𝑆 1 𝑥
𝜌 𝑚
1 𝐴 1 𝑥 EQ. D‐22
∙𝑢 ∙𝑆∙ 𝑥
𝜌 𝑚 𝑆
The holdup can be introduced in the right‐hand side term using the following equation:

223
 Appendix E: Derivation of Multiphase Flow Expressions M. Stanko

𝐴 𝐴 𝑚 1 1 EQ. D‐23
∙𝑢 ∙𝑆 ∙𝑢 ∙ 𝑥∙
𝑚 𝑚 𝑚 1 𝐻 ∙𝜌 1 𝐻 ∙𝜌

Expressing the liquid holdup in terms of Eq. D‐13, Eq. D‐23 can be rewritten as:

𝐴 1 1 EQ. D‐24
∙𝑢 ∙𝑆 𝑥∙ 𝑥∙ 𝑥∙𝜌
𝑚 𝑆∙𝜌 ∙ 1 𝑥 ∙𝜌
1 ∙𝜌 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥
𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥
𝐴 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 𝑥 𝑆∙ 1 𝑥 EQ. D‐25
∙𝑢 ∙𝑆
𝑚 𝜌 ∙𝜌 𝜌 𝜌

Substituting Eq. D‐25 in Eq. D‐22:

1 𝑥 𝑆∙ 1 𝑥 1 𝑥 EQ. D‐26
∙ 𝑥
𝜌 𝜌 𝜌 𝑆

KINETIC ENERGY‐AVERAGE MIXTURE DENSITY

𝑚 𝑢 𝑢 EQ. D‐27
𝑥∙ 1 𝑥 ∙
2∙𝜌 ∙𝐴 2 2
EQ. D‐28
1 2∙𝐴 𝑢 𝑢
∙ 𝑥∙ 1 𝑥 ∙
𝜌 𝑚 2 2
EQ. D‐29
1 𝐴 ∙𝑢 1 𝑥
∙ 𝑥
𝜌 𝑚 𝑆
EQ. D‐30
1 𝑚 1 𝑥
∙ 𝑥
𝜌 𝑚∙ 1 𝐻 ∙𝜌 𝑆
EQ. D‐31
1 𝑥 1 𝑥
∙ 𝑥
𝜌 1 𝐻 ∙𝜌 𝑆

Using Eq. D‐13, as a definition for the liquid holdup and substituting in Eq. D‐31:
EQ. D‐32
1 𝑥 1 𝑥
∙ 𝑥
𝜌 𝑆∙𝜌 ∙ 1 𝑥 𝑆
1 ∙𝜌
𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥
EQ. D‐33
1 𝑥 1 𝑥
∙ 𝑥
𝜌 𝑥∙𝜌 ∙𝜌 𝑆
𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥
EQ. D‐34
1 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 1 𝑥
∙ 𝑥
𝜌 𝜌 ∙𝜌 𝑆
EQ. D‐35
1 𝑥∙𝜌 𝑆∙𝜌 ∙ 1 𝑥 1 𝑥
∙ 𝑥
𝜌 𝜌 ∙𝜌 𝑆

224
Appendix E: Derivation of Multiphase Flow Expressions M. Stanko

EQ. D‐36
1 𝑥 𝑆∙ 1 𝑥 1 𝑥
∙ 𝑥
𝜌 𝜌 𝜌 𝑆

225
Appendix F: Oil & Gas processing diagrams M. Stanko

E. OIL & GAS PROCESSING DIAGRAMS

FIGURE E‐1. GAS PROCESSING FROM WELL TO SALES

226
Appendix F: Oil & Gas processing diagrams M. Stanko

FIGURE E‐2. GAS PROCESSING FROM WELL TO SALES (INCLUDING TYPICAL OPERATING VALUES)

227
 Appendix F: Derivation of local mass and volume fractions of oil gas and water as a function M. Stanko
of black oil properties

F. DERIVATION OF LOCAL MASS AND VOLUME FRACTIONS OF OIL GAS AND WATER AS A FUNCTION OF
BLACK OIL PROPERTIES

GAS, OIL AND WATER MASS FRACTIONS

Expressing the local gas volumetric rate in terms of the standard conditions rates and black oil properties:

𝐵 𝑅 ∙𝐵 EQ. F‐1
𝑞 𝑞 ∙ ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

The local mass flow rate of gas in terms of the local gas volumetric rate and the gas density:

𝐵 𝑅 ∙𝐵 EQ. F‐2
𝑚 𝑞 ∙𝜌 𝑞 ∙ ∙𝑞 ∙𝜌
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

Similarly for the oil and water mass rates:

Local oil volumetric rate:

𝐵 ∙𝑟 𝐵 EQ. F‐3
𝑞 𝑞 ∙ ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

Local mass rate of oil:

𝐵 ∙𝑟 𝐵 EQ. F‐4
𝑚 𝑞 ∙𝜌 𝑞 ∙ ∙𝑞 ∙𝜌
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

Local water volumetric rate

𝑞 𝐵 ∙𝑞 EQ. F‐5

Local mass rate of water:

𝑚 𝑞 ∙𝜌 𝐵 ∙𝑞 ∙𝜌 EQ. F‐6

The gas mass fraction is then expressed as:

𝑚 EQ. F‐7
𝑥
𝑚 𝑚 𝑚

Substituting Eq. F‐2, Eq. F‐4, Eq. F‐6 in Eq. F‐7 gives:

𝐵 𝑅 ∙𝐵
𝑞 ∙ ∙𝑞 ∙𝜌
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 EQ. F‐8
𝑥
𝐵 𝑅 ∙𝐵 𝐵 ∙𝑟 𝐵
𝑞 ∙ ∙𝑞 ∙𝜌 𝑞 ∙ ∙𝑞 ∙𝜌 𝐵 ∙𝑞 ∙𝜌
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

Simplifying terms:

𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞 ∙𝜌 EQ. F‐9
𝑥
𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞 ∙𝜌 𝑞 ∙𝐵 ∙𝑟 𝐵 ∙𝑞 ∙𝜌 1 𝑅 ∙𝑟 ∙𝐵 ∙𝑞 ∙𝜌

228
 Appendix F: Derivation of local mass and volume fractions of oil gas and water as a function M. Stanko
of black oil properties

𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞
𝑥 𝜌 𝜌 EQ. F‐10
𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞 𝑞 ∙𝐵 ∙𝑟 𝐵 ∙𝑞 ∙ 1 𝑅 ∙𝑟 ∙𝐵 ∙𝑞 ∙
𝜌 𝜌

Local densities of oil, gas and water are:

1 𝑟 EQ. F‐11
𝜌 ∙𝜌 ∙𝜌
𝐵 𝐵

𝑅 1 EQ. F‐12
𝜌 ∙𝜌 ∙𝜌
𝐵 𝐵

1 EQ. F‐13
𝜌 ∙𝜌
𝐵

The density ratios are:

𝑅 1
𝜌 ∙𝜌 ∙𝜌
𝐵 𝐵 EQ. F‐14
𝜌 1 𝑟
∙𝜌 ∙𝜌
𝐵 𝐵

1
𝜌 ∙𝜌
𝐵 EQ. F‐15
𝜌 1 𝑟
∙𝜌 ∙𝜌
𝐵 𝐵

Substituting Eq. F‐14 and Eq. F‐15 in Eq. F‐10 gives:

𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞
𝑥
𝑅 1 1 EQ. F‐16
∙𝜌 ∙𝜌 ∙𝜌
𝐵 𝐵 𝐵
𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞 𝑞 ∙𝐵 ∙𝑟 𝐵 ∙𝑞 ∙ 1 𝑟 1 𝑅 ∙𝑟 ∙𝐵 ∙𝑞 ∙ 1 𝑟
∙𝜌 ∙𝜌 ∙𝜌 ∙𝜌
𝐵 𝐵 𝐵 𝐵

Using:

𝑞 𝑅 ∙𝑞 EQ. F‐17

and

𝑊 EQ. F‐18
𝑞 ∙𝑞
1 𝑊

Gives:

𝑅 ∙𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞
𝑥
𝑅 1 1 EQ. F‐19
∙𝜌 ∙𝜌 𝑊 ∙𝜌
𝐵 𝐵 𝐵
𝑅 ∙𝑞 ∙𝐵 𝑅 ∙𝐵 ∙𝑞 𝑅 ∙𝑞 ∙𝐵 ∙𝑟 𝐵 ∙𝑞 ∙ 1 𝑟 1 𝑅 ∙𝑟 ∙𝐵 ∙1 𝑊 ∙𝑞 ∙ 1 𝑟
𝐵 ∙𝜌 𝐵 ∙𝜌 𝐵 ∙𝜌 𝐵 ∙𝜌

Simplifying terms:

229
 Appendix F: Derivation of local mass and volume fractions of oil gas and water as a function M. Stanko
of black oil properties

𝑅 ∙𝐵 𝑅 ∙𝐵
𝑥
𝑅 1
𝐵 ∙𝜌 𝐵 ∙𝜌 𝑊 𝜌 EQ. F‐20
𝑅 ∙𝐵 𝑅 ∙𝐵 𝑅 ∙𝐵 ∙𝑟 𝐵 ∙ 1 𝑟 1 𝑅 ∙𝑟 ∙ ∙ 1 𝑟
1 𝑊
𝐵 ∙𝜌 𝐵 ∙𝜌 𝐵 ∙𝜌 𝐵 ∙𝜌

Finally gives:

𝑅 𝑅
𝑥
𝑅 ∙𝜌 𝜌
EQ. F‐21
𝑊 𝜌
𝑅 𝑅 𝑅 ∙𝑟 1 ∙ 𝐵 1 𝑅 ∙𝑟 ∙ ∙ 𝐵
1 𝑊
𝜌 ∙𝑟 ∙𝜌 𝜌 ∙𝑟 ∙𝜌
𝐵 𝐵

For oil:

𝑅 ∙𝜌 𝜌
𝑅 ∙𝑟 1 ∙ 𝐵
𝜌 ∙𝑟 ∙𝜌
𝐵 EQ. F‐22
𝑥
𝑅 ∙𝜌 𝜌 𝑊 𝜌
𝑅 𝑅 𝑅 ∙𝑟 1 ∙ 𝐵 1 𝑅 ∙𝑟 ∙ ∙ 𝐵
1 𝑊
𝜌 ∙𝑟 ∙𝜌 𝜌 ∙𝑟 ∙𝜌
𝐵 𝐵

For water:

𝑊 𝜌
1 𝑅 ∙𝑟 ∙ ∙ 𝐵
1 𝑊
𝜌 ∙𝑟 ∙𝜌
𝐵 EQ. F‐23
𝑥
𝑅 ∙𝜌 𝜌 𝑊 𝜌
𝑅 𝑅 𝑅 ∙𝑟 1 ∙ 𝐵 1 𝑅 ∙𝑟 ∙ ∙ 𝐵
1 𝑊
𝜌 ∙𝑟 ∙𝜌 𝜌 ∙𝑟 ∙𝜌
𝐵 𝐵

If one assumes 𝑟 0, then the expressions are simplified as follows:

𝑅 𝑅
𝑥 EQ. F‐24
𝜌 𝑊 𝜌
𝑅 ∙
𝜌 1 𝑊 𝜌

𝜌
𝑅
𝜌 EQ. F‐25
𝑥
𝜌 𝑊 𝜌
𝑅 ∙
𝜌 1 𝑊 𝜌

𝑊 𝜌
∙
1 𝑊 𝜌 EQ. F‐26
𝑥
𝜌 𝑊 𝜌
𝑅 ∙
𝜌 1 𝑊 𝜌

230
 Appendix F: Derivation of local mass and volume fractions of oil gas and water as a function M. Stanko
of black oil properties

GAS, OIL AND WATER VOLUME FRACTIONS

The gas volume fraction is defined as:

𝑞 EQ. F‐27
𝛼
𝑞 𝑞 𝑞

Expressing local volumetric rates of gas, oil and water in terms of standard conditions rates and BO properties:

𝐵 𝑅 ∙𝐵
𝑞 ∙ ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 EQ. F‐28
𝛼
𝐵 𝑅 ∙𝐵 𝐵 ∙𝑟 𝐵
𝑞 ∙ ∙𝑞 𝑞 ∙ ∙𝑞 𝐵 ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

Using:

𝑞 𝑅 ∙𝑞 EQ. F‐29

and

𝑊 EQ. F‐30
𝑞 ∙𝑞
1 𝑊

And simplifying, gives:

𝑅 𝑅
𝛼 EQ. F‐31
𝐵 𝑊 𝐵
𝑅 𝑅 ∙ 𝑅 ∙𝑟 1 1 𝑅 ∙𝑟 ∙ ∙
𝐵 1 𝑊 𝐵

For oil:

𝐵
∙ 𝑅 ∙𝑟 1
𝐵 EQ. F‐32
𝛼
𝐵 𝑊 𝐵
𝑅 𝑅 ∙ 𝑅 ∙𝑟 1 1 𝑅 ∙𝑟 ∙ ∙
𝐵 1 𝑊 𝐵

For water

𝑊 𝐵
1 𝑅 ∙𝑟 ∙ ∙
1 𝑊 𝐵 EQ. F‐33
𝛼
𝐵 𝑊 𝐵
𝑅 𝑅 ∙ 𝑅 ∙𝑟 1 1 𝑅 ∙𝑟 ∙ ∙
𝐵 1 𝑊 𝐵

If one assumes 𝑟 0, then the expressions are simplified as follows:

𝑅 𝑅
𝛼 EQ. F‐34
𝐵 𝑊 𝐵
𝑅 𝑅 ∙
𝐵 1 𝑊 𝐵

231
 Appendix F: Derivation of local mass and volume fractions of oil gas and water as a function M. Stanko
of black oil properties

𝐵
𝐵 EQ. F‐35
𝛼
𝐵 𝑊 𝐵
𝑅 𝑅 ∙
𝐵 1 𝑊 𝐵

𝑊 𝐵
∙
1 𝑊 𝐵 EQ. F‐36
𝛼
𝐵 𝑊 𝐵
𝑅 𝑅 ∙
𝐵 1 𝑊 𝐵

The volume fraction of oil in liquid:

𝐵 ∙𝑟 𝐵
𝑞 𝑞 ∙ ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 EQ. F‐37
𝛼 ,
𝑞 𝑞 𝐵 ∙𝑟 𝐵
𝑞 ∙ ∙𝑞 𝐵 ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟

Using:

𝑞 𝑅 ∙𝑞 EQ. F‐38

and

𝑊 EQ. F‐39
𝑞 ∙𝑞
1 𝑊

Gives:

𝐵 ∙𝑟 𝐵
𝑅 ∙𝑞 ∙ ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 EQ. F‐40
𝛼 , 𝐵 ∙𝑟 𝐵 𝑊
𝑅 ∙𝑞 ∙ ∙𝑞 𝐵 ∙ ∙𝑞
1 𝑅 ∙𝑟 1 𝑅 ∙𝑟 1 𝑊

Simplifying terms:

𝑅 ∙ 𝐵 ∙𝑟 𝐵
𝛼 , EQ. F‐41
𝑊
𝑅 ∙ 𝐵 ∙𝑟 𝐵 1 𝑅 ∙𝑟 ∙𝐵 ∙
1 𝑊

Gives finally:

𝑅 ∙𝑟 1
𝛼 , EQ. F‐42
𝐵 𝑊
𝑅 ∙𝑟 1 1 𝑅 ∙𝑟 ∙ ∙
𝐵 1 𝑊

If one assumes 𝑟 0, then the expression is simplified as follows:

𝐵
𝛼 , EQ. F‐43
𝑊
𝐵 𝐵 ∙
1 𝑊

232
 Appendix F: Derivation of local mass and volume fractions of oil gas and water as a function M. Stanko
of black oil properties

Liquid density for a mixture of oil and water (assuming 𝑟 0):

𝐵 𝐵 EQ. F‐44
𝜌 ∙𝜌 ∙𝜌
𝑊 𝑊
𝐵 𝐵 ∙ 𝐵 𝐵 ∙
1 𝑊 1 𝑊

Using Eq. F‐12 and Eq. F‐13 gives:

𝑊
𝐵 𝑅 1 𝐵 ∙ 1
1
𝑊 EQ. F‐45
𝜌 ∙ ∙𝜌 ∙𝜌 ∙ ∙𝜌
𝑊 𝐵 𝐵 𝑊 𝐵
𝐵 𝐵 ∙ 𝐵 𝐵 ∙
1 𝑊 1 𝑊

simplifying:

𝑊
𝑅 ∙𝜌 𝜌 𝜌 ∙
1 𝑊 EQ. F‐46
𝜌
𝑊
𝐵 𝐵 ∙
1 𝑊

233
 Appendix G: Derivation of the expression of field producing gas‐oil ratio M. Stanko

G. DERIVATION OF THE EXPRESSION OF FIELD PRODUCING GAS‐OIL RATIO

The producing gas oil ratio (𝑅 ) can be expressed as the ratio between gas rate and oil rate:
𝑞 𝑞
𝑅 EQ. G‐1
𝑞 𝑞

Neglecting the oil condensing from the gas (𝑞 0):

𝑞 𝑞 𝑞 𝑞 EQ. G‐2
𝑅
𝑞 𝑞 𝑞

The second term in the equation is the solution gas oil ratio (𝑅 ). The first term can be expressed as a local rate
of oil or gas times the oil volume factor and the gas volume factor.
𝑞
𝐵 EQ. G‐3
𝑅 𝑞 𝑅
𝐵

The local flow rates depend on the IPR of each phase (and integrating the pressure function from pwf to pR).
However, as a simplification, if one considers the reservoir as a tank with uniform pressure pR then the integral
and well geometric effects disappear, and each rate will be proportional to kr/μ. Therefore:
𝑘
𝜇 𝐵 EQ. G‐4
𝑅 ⎛ ⎞ 𝑅
𝑘
⎝𝜇 𝐵 ⎠

234
 Appendix H: Gas lift optimization M. Stanko

H. GAS LIFT OPTIMIZATION

Example 1: Single well, unconstrained maximization of economic revenue by adjusting gas injection rate
A simple but very typical revenue function is defined by Eq. H‐1:

𝑓 𝑞 𝑞 ∙𝑃 𝑞 ∙𝑃 𝑓 𝑞 ∙𝑃 𝑞 ∙𝑃 EQ. H‐1
, , , ,

Where:
𝑃 Oil price [USD/stb/d, USD/Sm3/d]
𝑃 Cost of injection gas [USD/MMSCFD, USD/MSm3/d]
𝑓 Revenue function

Deriving the function with respect to the adjustable variable:

𝑑𝑓 𝑞 , 𝑑𝑓 𝑞 , 𝑑𝑓 𝑞 , 𝑃 EQ. H‐2
∙ 𝑃 𝑃 0⇒
𝑑𝑞 , 𝑑𝑞 , 𝑑𝑞 , 𝑃

The maximum revenue is therefore achieved for the point in the gas lift performance curve where the
derivative is exactly equal to the ratio between the injection gas cost and the oil price. In general, the gas price
is much smaller than the oil price, yielding that the derivative must be very close to zero.
Example 2: Single well, maximization of oil production with limited gas injection rate available
To include the limitation on injection gas available (qg,inj ≤ qg,inj TOT) the Lagrange function[3‐2] is created:
EQ. H‐3
𝐿 𝑞 , 𝑓 𝑞 , 𝜆∙ 𝑞 , 𝑞 ,

The maximum is given when the derivative of the function with respect to the adjustable variable is equal to
zero (Eq. H‐4) and when the additional conditions (Eq. H‐5, Eq. H‐6, Eq. H‐7) are met:

𝑑𝐿 𝑞 , 𝑑𝑓 𝑞 , 𝑑𝑓 𝑞 , EQ. H‐4
𝜆 0⇒ 𝜆
𝑑𝑞 , 𝑑𝑞 , 𝑑𝑞 ,
EQ. H‐5
𝜆∙ 𝑞 , 𝑞 , 0
EQ. H‐6
𝜆 0
EQ. H‐7
𝑞 , 𝑞 , 0

There are two possible solutions:

Solution 𝑑𝑓 𝑞 , Valid only if there is enough gas available (qg,inj ≤ qg,inj TOT)
𝜆 0, 0
1: 𝑑𝑞 ,
Solution 𝑑𝑓 𝑞 , Valid only if all the gas available is used (qg,inj = qg,inj TOT). Please note
𝜆 0, 𝜆
2: 𝑑𝑞 , that due to the condition (Eq. H‐6) on lambda, the derivative MUST
NOT be negative.

Example 3: Single well, maximization of revenue with limited gas injection rate.
The Lagrange function is created:
EQ. H‐8
𝐿 𝑞 , 𝑓 𝑞 , 𝜆∙ 𝑞 , 𝑞 ,

235
 Appendix H: Gas lift optimization M. Stanko

The maximum is given when the derivative of the function with respect to the adjustable variable is equal to
zero (Eq. H‐9) and when the additional conditions (Eq. H‐10, Eq. H‐11, Eq. H‐12) are met:

𝜕𝐿 𝑞 , 𝑑𝑓 𝑞 , 𝑑𝑓 𝑞 , 𝑑𝑓 𝑞 , 𝜆 𝑃 EQ. H‐9
𝜆 0⇒ ∙ 𝑃 𝑃 𝜆⇒
𝜕𝑞 , 𝑑𝑞 , 𝑑𝑞 , 𝑑𝑞 , 𝑃
EQ. H‐10
𝜆∙ 𝑞 , 𝑞 , 0
EQ. H‐11
𝜆 0
EQ. H‐12
𝑞 , 𝑞 , 0

There are two possible solutions:

Solution 𝑑𝑓 𝑞 , 𝑃 Valid only if there is enough gas available (qg,inj ≤ qg,inj TOT)
𝜆 0,
1: 𝑑𝑞 , 𝑃
Solution 𝑑𝑓 𝑞 , 𝜆 𝑃 Valid only if all the gas available is used (qg,inj = qg,inj TOT).
𝜆 0,
2: 𝑑𝑞 , 𝑃

Example 4: unconstrained oil production maximization on a group of wells by adjusting the individual well gas
lift injection rate
The present development assumes that the operation of an individual well is independent from the others
(the operating wellhead pressure remains constant despite of the operating conditions of the individual wells).
This is because the mathematical procedure employed requires the objective function (e.g. total oil
production) to be additively separable (a function that can expressed as the summation of two or more
functions each one depending on only one variable).
The total oil production function (F) is the sum of the individual (i) well oil production (fi). The total number of
wells is N.

EQ. H‐13
𝐹 𝑞 , ,𝑞 , ,𝑞 , ,… 𝑓 𝑞 ,

F is a multivariate (N) additively separable scalar function. A necessary condition for this function to be
maximum is that the elements of its gradient must be equal to zero:

𝜕𝐹 𝑞 , ,𝑞 , ,𝑞 , ,… 𝜕𝑓 𝑞 , EQ. H‐14
0
𝜕𝑞 , 𝜕𝑞 ,

The maximum of the compound oil production is when all the oil production of the individual wells is also
maximum.
Example 5: Revenue maximization on a group of wells by adjusting the gas lift injection rate
The total revenue function (frevTOT) is the sum of the individual well oil production (frev,i). The total number of
wells is N.

EQ. H‐15
𝑓 𝑞 , ,𝑞 , ,𝑞 , ,… 𝑓 , 𝑞 ,

236
 Appendix H: Gas lift optimization M. Stanko

frevTOT is a multivariate (N) additively separable scalar function. A necessary condition for this function to be
maximum is that the elements of its gradient must be equal to zero:

𝜕𝑓 𝑞 , ,𝑞 , ,𝑞 , ,… 𝜕𝑓 , 𝑞 , 𝜕𝑓 𝑞 , EQ. H‐16
∙ 𝑃 𝑃 0
𝜕𝑞 , 𝜕𝑞 , 𝜕𝑞 ,

The maximum of the compound revenue is achieved when all the revenues of the individual wells are also
maximum.
Example 6: revenue maximization of a group of wells by adjusting the gas lift injection rate with limited gas
The total revenue function (frevTOT) is the sum of the individual well oil production (frev,i). The total number of
wells is N.

EQ. H‐17
𝑓 𝑞 , ,𝑞 , ,𝑞 , ,… 𝑓 , 𝑞 ,

frevTOT is a multivariate (N) additively separable scalar function. In order to include the limitation on injection
gas available (∑qg,inj < qg,inj TOT) the method of Lagrange multipliers is used. The Lagrange function is created:

EQ. H‐18
𝐿 𝑞 , ,𝑞 , ,𝑞 , ,… 𝑓 , 𝑞 , 𝜆∙ 𝑞 , 𝑞 ,

A necessary condition for this function to be maximum is that the elements of its gradient must be equal to
zero (Eq. H‐19) and when the additional conditions (Eq. H‐20, Eq. H‐21, Eq. H‐22) are met:

𝜕𝐿 𝑞 , ,𝑞 , ,𝑞 , ,… 𝜕𝑓 , 𝑞 , 𝜕𝑓 𝑞 , 𝜆 𝑃 EQ. H‐19
𝜆 0⇒
𝜕𝑞 , 𝜕𝑞 , 𝜕 𝑃

EQ. H‐20
𝜆∙ 𝑞 , 𝑞 , 0

EQ. H‐21
𝜆 0

EQ. H‐22
𝑞 , 𝑞 ,

There are two possible solutions:

Solution 𝜕𝑓 𝑞 , 𝑃 All the wells are operating in their maximum revenue point.
𝜆 0,
1: 𝜕𝑞 , 𝑃 Valid only if there is enough gas available (∑qg,inj < qg,inj TOT)
Solution 𝜕𝑓 𝑞 , 𝜆 𝑃 All wells are operating at the same derivative in the gas lift
𝜆 0,
2: 𝜕𝑞 , 𝑃 performance curve. Valid only if all the gas available is used
(∑qg,inj ‐ qg,inj TOT = 0)

The procedure described is also applicable for any situation where the well (or group of wells) has a concave
performance curve of oil production vs an adjustable parameter. For example, ESP lifted wells with diluent
injection at the ESP suction also exhibit a similar performance curve.

237
 Appendix I: Analytical expression of revenue NPV considering linear oil price variation M. Stanko

I. ANALYTICAL EXPRESSION OF REVENUE NPV CONSIDERING A LINEAR DIMENSIONLESS PRODUCTION

POTENTIAL, CONTINUOUS DISCOUNTING AND OIL PRICE VARYING LINEARLY IN TIME. CASE STUDY: OIL
OFFSHORE FIELD.

Assume the oil price shows a linear behavior with time:

𝑃 𝑡 𝑃 𝑚 ∙𝑡 EQ. I‐1

This gives the following expression for the revenue NPV:

𝑁𝑃𝑉 𝑞 , ∙𝑃 ∙𝑡
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆
𝑒 𝑒 𝑒 𝑒
∙
𝑖 𝑚 𝑖
∆ EQ. I‐2
∙
𝑡 ∙ 𝑞 , ∙𝑚 ∙𝑡∙𝑒 𝑑𝑡 𝑡

∙ ∆ ∙
∙ 𝑞 , ∙𝑒 ∙𝑚 ∙𝑡∙𝑒 𝑑𝑡
∆

Solving the second‐to‐last integral and factorizing the last integral:

𝑁𝑃𝑉 𝑞 , ∙𝑃 ∙𝑡
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆
𝑒 𝑒 𝑒 𝑒
∙
𝑖 𝑚 𝑖
∙ ∆ EQ. I‐3
𝑒 1
𝑞 , ∙𝑚 ∙𝑡 ∙ 𝑡 𝑞 , ∙𝑚 ∙𝑡
𝑖 𝑖

∙ ∆ ∙
∙𝑒 ∙ 𝑒 ∙ 𝑡 ∙ 𝑑𝑡
∆

Solving the last integral and evaluating the second‐to‐last integral:

𝑁𝑃𝑉 𝑞 , ∙𝑃 ∙𝑡
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆
𝑒 𝑒 𝑒 𝑒
∙
𝑖 𝑚 𝑖
𝑞 , ∙𝑚 ∙𝑡
EQ. I‐4
∙ ∆ ∙
𝑒 1 𝑒 1
∙ 𝑡 ∆𝑡 𝑡 𝑞 , ∙
𝑖 𝑖 𝑖 𝑖
∙
∙ ∆
𝑒 1
∙𝑡 ∙𝑒 ∙ 𝑡
𝑚 𝑖 𝑚 𝑖 ∆

Evaluating the last integral:

238
Appendix I: Analytical expression of revenue NPV considering linear oil price variation M. Stanko

𝑁𝑃𝑉 𝑞 , ∙𝑃 ∙𝑡
∙ ∙ ∆ ∙ ∆ ∙ ∙ ∆
𝑒 𝑒 𝑒 𝑒
∙
𝑖 𝑚 𝑖
𝑞 , ∙𝑚 ∙𝑡
𝑒 ∙ ∆
1 𝑒 ∙
1 EQ. I‐5
∙ 𝑡 ∆𝑡 𝑡 𝑞 , ∙
𝑖 𝑖 𝑖 𝑖
∙ ∆
∙𝑡 ∙𝑒
∙ ∙ ∆
𝑒 1 𝑒 1
∙ 𝑡 𝑡 ∆𝑡
𝑚 𝑖 𝑚 𝑖 𝑚 𝑖 𝑚 𝑖

239
 Appendix J: Some Style Comments for Technical Communication M. Stanko

J. SOME STYLE COMMENTS FOR TECHNICAL COMMUNICATION (PARAPHRASING THE NOTES OF M.

STANDING AND M. GOLAN)
 Introduce smoothly the topic to the reader. Use as many aids as possible to accomplish that.
 Think of what you want to communicate with the sentence before writing it.
 Use active verbs.
 Give the most important information at the beginning of the sentence. It must be clear and objective.
 Differentiate between “observed”, “calculated”, “assumed”, “guessed”.
 Limit sentences to 30 words.
 Avoid clichés, watery and loaded statements that do not add any important information.
 Reference properly your statements.
 Give the proper context and take enough time when introducing and discussing a figure or a diagram.
 All figures and tables should be discussed in the text.
 Use proper English, select the right words and verbs.
 Your statements should be, as much as possible, defendable in a court of Law.

240

View publication stats