132
Production Reallocation of the Pazflor Oil Field Correction of the Produced Volumes of Oil, Water and Gas per Well using the Integrated Field Management (IFM) by Petex Yolanda da Purificação Mambo Gaspar Tati Dissertation for obtention of the Master Degree in Petroleum Engineering Supervisors: Prof. Dr. Amílcar de Oliveira Soares (IST), Eng. Sebastien Pérrier (TOTAL) Júri President: Prof. Dr. Maria João Correia Colunas Pereira Supervisor: Prof. Dr. Amílcar de Oliveira Soares Vogais: Prof. Dr. Leonardo Azevedo Guerra Raposo Pereira September 2015

Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

Production Reallocation of the Pazflor Oil Field

Correction of the Produced Volumes of Oil, Water and Gas per Well using the Integrated Field Management (IFM) by Petex

Yolanda da Purificação Mambo Gaspar Tati

Dissertation for obtention of the Master Degree in

Petroleum Engineering

Supervisors: Prof. Dr. Amílcar de Oliveira Soares (IST),

Eng. Sebastien Pérrier (TOTAL)

Júri

President: Prof. Dr. Maria João Correia Colunas Pereira

Supervisor: Prof. Dr. Amílcar de Oliveira Soares

Vogais: Prof. Dr. Leonardo Azevedo Guerra Raposo Pereira

September 2015

Page 2: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

2

Page 3: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

3

Thanks

To all of those who participated in my education. First, to my father, who taught me not to fail. Second,

to my mother, who taught me to believe. To Camila Gonçalves and Mr. Fialho, who taught me there

was also a place for me. To all the teachers that gave me the best part of them and encouraged me to

be the best part of me. To all the colleagues I’ve always wanted to be around of. To Rafa, that was

always there. And, finnally, to Carlos who was heaven sent.

Page 4: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

4

Page 5: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

5

Abstract

The present study aims to allocate the volumes of oil, water and gas to the different wells and

producing lines of the Pazflor oil field.

The production allocation was held through the measurement and evaluation of the deviations first

obtained by the readings of the multiphase flow rates.

In the Pazflor producing field, the measurement of the produced volumes of oil, water and gas is first

held at the well level by multiphase flow meters. These measurements are associated to different

ranges of uncertainties due to technical limitations of this equipment.

The correction proposed for the Multiphase Flow Meters readings is obtained through the Integrated

Field Manager. The corrections on the production curves are done by modifying the Watercut and the

Gas to Oil Ratio on the periods in which the Bottom Hole Pressure estimated deviates in more than 3

bar from the real Bottom Hole Pressure (directly measured through the Down Hole Gauge). These are

the only two parameters that are altered in order to obtain new oil, water and gas rates that are

consistent with the directly measured Bottom Hole Pressure.

Here, density and friction assumptions are associated to the volumes recorded by the Multiphase Flow

Meters and are used to compute a Bottom Hole Pressure and a Tubing Head Pressure that has to

match the directly measured pressure obtained by the Down Hole Gauge in order to be considered as

valid. When out of the admissible range of pressure, the correspondent Watercut and GOR curves

were corrected in order to match the behavior suggested by the Bottom Hole Pressure. The

corrections and adjustments follow the fundamental physical constraints that affect the reservoir and

are always done with maximum parsimony.

The Multiphase Flow Meters measurements are based on density functions and on the different

sensors readings that can automatically be switched to either an oil continuous mode- recording the

permittivity- or a water continuous mode- recording the conductivity. It switches to the water

continuous mode if the watercut is higher than 45%. It often occurs that the conductivity of reference

which the water continuous mode uses to calculate does no longer matches the conductivity of the

produced water in the medium. The reliability on the Multiphase Flow Meter measurements decrease

on the group of wells that are linked to a higher production of water, where water can be misread as

hydrocarbons or vice versa.

For the cases where the Watercut is higher than 45%, two scenarios were created in order to

determine a range of uncertainty on admissible values given by the Integrated Field Manager.

The well physical parameters evolution in time was also followed, mainly in cases where strong

corrections had to be applied in order to obtain matching pressure values. The density of the flow (also

related to the permittivity and conductivity) is used to confirm the abrupt changes in the different

phases components. The Well Head Pressure, the Well Head Temperature and the Bottom Hole

Pressure are also used to infer about the realiability of the corrected production curves.

Page 6: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

6

These Multiphase Flow Meters corrections lead to a new Multiphase Flow Meters production scenario.

In order to infer about the global admissibility of this process and of the IFM tool, a process of data

reconciliation for validation took place. The total of the produced volumes of oil, water and gas was

determined at the levels of the Subsea Separation Units, Pumps and Topside through the

measurements held by the flow transmitters installed in the producing lines and Floating Production

Storage Offloading unit.

Through a closed loop process, it is possible to take global conclusions about the precisions of the

measurements at different production stages and to infer about the reliability of the different sources of

data acquisition used nowadays in this field. The flow rate value registered by the oil tanks flow

transmitters (before selling the oil) is considered to have no uncertainty associated to it. By comparing

all the oil production curves with the last one (oil tanks) it is possible to go back on the process and

improve the corrections applied. The periods of flow rate mismatch in between the wells production

curves and the topside (oil tanks) would directly allow to identify the producing line showing more

discrepancies for the identic production period as the degrees of freedom would be reduced at this

stage. After that, it would be possible to more accurately spot the wells in which the flow rates of oil,

water and gas had to be corrected in order to match the real flow rate measured before selling the oil.

The empirical method that was applied showed no biased results. This robust empirical method

allowed to estimate the produced volumes of oil, water and gas with an estimation error that is not

higher than approximately 5% during more than 80% of the studied period. It was also possible to

discover that the Multiphase Flow Meter was underestimating in more than 50% the produced water

and gas during 90% of the wells producing time. For oil, the Multiphase Flow Meter shows a maximum

error on estimation of approximately 25% during 70% of the producing period.

This deviations openned a path to question the reliability on the models that are currently on use. For

some wells, the building consistent production scenarios implied allocating Gas to Oil Ratio values

bellow the Rs or a little displacement on time of the water breakthrough for other wells.

This consisted in a considerable adding value conclusions and acknowledgements since it brought

attention to where should resources be applied in order to achieve the main goal: producing oil.

This correction process was applied to all the Pazflor wells since the beginning of production (1st

August 2011) until 25th February 2015 so the main evolution trend of the production was captured.

Key-words: Production Reallocation; Multiphase Flow Metering; Integrated Field Management;

Pazflor; TOTAL; Angola; Deep Offshore.

Page 7: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

7

Table of Contents

THANKS ............................................................................................................................... 3

ABSTRACT .......................................................................................................................... 5

TABLE OF CONTENTS ........................................................................................................ 7

TABLE OF FIGURES............................................................................................................ 9

LIST OF ACRONYMS ......................................................................................................... 15

CHAPTER 1. INTRODUCTION .......................................................................................... 16

CHAPTER 2. TOTAL'S BLOCK 17 .................................................................................... 18

2.1 BLOCK 17 – A WORLD SCALE BLOCK ....................................................................... 18

CHAPTER 3. PAZFLOR ..................................................................................................... 19

3.1 A TOP OF THE ART PROJECT .................................................................................... 19

3.2 STRUCTURE AND EQUIPMENT ................................................................................... 19

3.3 MEASUREMENT OF PRODUCED VOLUMES .................................................................. 21

3.3.1 The Multiphase Flow Meters (MPFM) ............................................................... 21

CHAPTER 4. PRODUCTION ALLOCATION ...................................................................... 24

4.1 THE NEED AND THE GOAL ......................................................................................... 24

4.2 THE SOFTWARE - INTEGRATED PRODUCTION MODELLING (IPM) BY PETROLEUM EXPERTS

24

4.3 IFM PROCESS - CORRECTION METHODOLOGY ........................................................... 26

4.4 IFM PROCESS - THE DARCY'S LAW AS THE RELATION IN BETWEEN GOR AND WATERCUT

WITH PRESSURE ................................................................................................................ 28

4.5 CORRECTION CRITERIA FOR THE REALLOCATION ....................................................... 30

4.5.1 The Gas-Oil Ratio minimum value .................................................................... 30

4.5.2 Maximum Parsimony Criteria ........................................................................... 31

4.5.3 Reservoir Depletion in time .............................................................................. 32

4.6 THE CORRECTION: MPFM RAW VS MPFM CORRECTED ............................................. 33

4.6.1 Production line: P10 ......................................................................................... 33

4.6.2 Production line - P20 ........................................................................................ 50

4.6.3 Production line - P30 ........................................................................................ 56

4.6.4 Production line - P40 ........................................................................................ 58

CHAPTER 5. METHOD ROBUSTNESS ............................................................................. 60

Page 8: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

8

CHAPTER 6. DATA RECONCILIATION RESULTS ........................................................... 61

CHAPTER 7. UNCERTAINTIES AND ADMISSIBILITY ..................................................... 63

CHAPTER 8. CONCLUSIONS ........................................................................................... 77

CHAPTER 9. WAY FORWARD .......................................................................................... 78

CHAPTER 10. REFERENCES ............................................................................................ 79

ANNEXES ........................................................................................................................... 80

ANNEX I. CUMULATIVE FREQUENCY GRAPHS ....................................................................... 80

ANNEX II. DEVIATION GRAPHS .......................................................................................... 104

ANNEX III. CURVE REPAIR + MOVING AVERAGES ALGORITHM ............................................. 125

ANNEX IV. RS TABLE FOR THE PAZFLOR WELLS ................................................................. 132

Page 9: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

9

Table of Figures

FIGURE 1 PRODUCTION REALLOCATION WORKFLOW SCHEME................................................. 17

FIGURE 2 PAZFLOR PRODUCING LINES SCHEME. ........................................................................ 20

FIGURE 3 EXAMPLE OF MPFM'S AND ILLUSTRATION OF ITS LOCATION SUBSEA. ................... 21

FIGURE 4 SIMPLIFIED ILLUSTRATION OF THE VENTURI TUBE. ................................................... 22

FIGURE 5 THE MPFM MEASURING PRINCIPLES. ............................................................................ 22

FIGURE 6 IFM AND MODELS INTERACTION. ................................................................................... 25

FIGURE 7 CALIBRATION IN THE ALLOCATION WORKFLOW. ........................................................ 26

FIGURE 8 REBUILDING PRODUCTION HISTORY FOR ZNA-ED WELL. ......................................... 27

FIGURE 9 GRAVITY AND FRICTION IN A FLOWING WELL.FOR THE SAME QTOTAL, THE

PRESSURE DROP INCREASES WHEN THERE IS MORE WATER PRESENT IN THE FLOW. 29

FIGURE 10 EXAMPLE OF TWO POSSIBLE CORRECTED SCENARIOS FOR PRP-FB-A. .............. 31

FIGURE 11 GRAPHIC ILLUSTRATION OF THE MINIMUM CHANGE CRITERIA APPLIED. ............ 32

FIGURE 12 GRAPHIC ILLUSTRATION OF THE CRITERIA FOR THE LEVEL OF CERTAINTY

ABOUT CORRECTIONS AT A CERTAIN STAGE OF THE PRODUCING TIME. ........................ 33

FIGURE 13 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE LINE P10. .................................................................................. 34

FIGURE 14 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE LINE P10. .............................................................................................. 34

FIGURE 15 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE WELL ZNA-E0A ......................................................................... 35

FIGURE 16 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE WELL ZNA-E0A..................................................................................... 36

FIGURE 17 WELL HEAD TEMPERATURE EVOLUTION IN TIME FOR THE WELL ZNA-E0A ......... 36

FIGURE 18 WELL HEAD PRESSURE EVOLUTION IN TIME FOR THE WELL ZNA-E0A ................. 37

FIGURE 19 BOTTOM-HOLE PRESSURE EVOLUTION IN TIME FOR THE WELL ZNA-E0A ........... 37

FIGURE 20 EVOLUTION OF THE PRESSURE AT THE WELL HEAD. P1 – MEASURED

PRESSURE. P2 – ESTIMATED PRESSURE. .............................................................................. 38

FIGURE 21 P1 VS P2 – LINEAR REGRESSION. P1 – MEASURED PRESSURE. P2 – ESTIMATED

PRESSURE. ................................................................................................................................... 38

FIGURE 22 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE WELL ZNA-E0E. ........................................................................ 39

FIGURE 23 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE WELL ZNA-E0E..................................................................................... 40

FIGURE 24 WELL HEAD TEMPERATURE EVOLUTION IN TIME FOR THE WELL ZNA-E0E ......... 40

FIGURE 25 WELL HEAD PRESSURE EVOLUTION IN TIME FOR THE WELL ZNA-E0E ................. 41

FIGURE 26 BOTTOM-HOLE PRESSURE EVOLUTION IN TIME FOR THE WELL ZNA-E0E ........... 41

FIGURE 27 EVOLUTION OF THE PRESSURE AT THE WELL HEAD. P1 – MEASURED

PRESSURE. P2 – ESTIMATED PRESSURE. .............................................................................. 42

Page 10: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

10

FIGURE 28 P1 VS P2 – LINEAR REGRESSION. P1 – MEASURED PRESSURE. P2 – ESTIMATED

PRESSURE. ................................................................................................................................... 42

FIGURE 29 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE WELL ZNA-E0D. ........................................................................ 43

FIGURE 30 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE WELL ZNA-E0D .................................................................................... 44

FIGURE 31 WELL HEAD PRESSURE EVOLUTION IN TIME FOR THE WELL ZNA-E0D ................. 44

FIGURE 32 WELL HEAD TEMPERATURE EVOLUTION IN TIME FOR THE WELL ZNA-E0D ......... 45

FIGURE 33 BOTTOM-HOLE PRESSURE EVOLUTION IN TIME FOR THE WELL ZNA-E0D ........... 45

FIGURE 34 EVOLUTION OF THE PRESSURE AT THE WELL HEAD. P1 – MEASURED

PRESSURE. P2 – ESTIMATED PRESSURE. .............................................................................. 46

FIGURE 35 P1 VS P2 – LINEAR REGRESSION. P1 – MEASURED PRESSURE. P2 – ESTIMATED

PRESSURE. ................................................................................................................................... 46

FIGURE 36 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE WELL PRP-F0BA ...................................................................... 47

FIGURE 37 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE WELL PRP-F0BA .................................................................................. 48

FIGURE 38 WELL HEAD TEMPERATURE EVOLUTION IN TIME FOR THE WELL PRP-F0BA ....... 48

FIGURE 39 WELL HEAD PRESSURE EVOLUTION IN TIME FOR THE WELL PRP-F0BA .............. 49

FIGURE 40 BOTTOM-HOLE PRESSURE IN TIME FOR THE WELL PRP-F0BA ............................... 49

FIGURE 41 BOTTOM-HOLE PRESSURE EVOLUTION IN TIME FOR THE WELL PRP-F0BA ......... 50

FIGURE 42 P1 VS P2 – LINEAR REGRESSION. P1 – MEASURED PRESSURE. P2 – ESTIMATED

PRESSURE. ................................................................................................................................... 50

FIGURE 43 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE LINE P20. .................................................................................. 51

FIGURE 44 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE LINE P20. .............................................................................................. 52

FIGURE 45 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE WELL PRP-FA0. ........................................................................ 53

FIGURE 46 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR WELL PRP-FA0. ............................................................................................ 53

FIGURE 47 WELL HEAD TEMPERATURE EVOLUTION IN TIME FOR THE WELL PRP-FA0 ......... 54

FIGURE 48 WELL HEAD PRESSURE EVOLUTION IN TIME FOR THE WELL PRP-FA0 ................. 54

FIGURE 49 WELL HEAD PRESSURE EVOLUTION IN TIME FOR THE WELL PRP-FA0 ................. 55

FIGURE 50 P1 VS P2 ........................................................................................................................... 55

FIGURE 51 EVOLUTION OF THE PRESSURE AT THE WELL HEAD. .............................................. 56

FIGURE 52 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE LINE P30. .................................................................................. 57

FIGURE 53 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE LINE P30. .............................................................................................. 57

Page 11: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

11

FIGURE 54 COMPARISON OF THE WATER AND OIL FLUX FOR THE MPFM RAW AND THE

MPFM CORRECTED FOR THE LINE P40. .................................................................................. 59

FIGURE 55 COMPARISON OF THE GAS FLUX FOR THE MPFM RAW AND THE MPFM

CORRECTED FOR THE LINE P40. .............................................................................................. 59

FIGURE 56 ABOVE AND BELOW, RESPECTIVELY, THE RESUME OF THE PRODUCTION

SUBSEA NETWORK AND OF THE TOPSIDE STORAGE AND OFFLOADING NETWORK. ..... 62

FIGURE 57 COMPARISON OF OIL PRODUCTION IN ORDER OF A MEASURED TIME. ................ 63

FIGURE 58 COMPARISON OF WATER PRODUCTION IN ORDER OF A MEASURED TIME. ........ 64

FIGURE 59 COMPARISON OF GAS PRODUCTION IN ORDER OF A MEASURED TIME. .............. 64

FIGURE 60 CUMULATIVE VIEW FOR OIL. ......................................................................................... 65

FIGURE 61 CUMULATIVE VIEW FOR WATER. .................................................................................. 65

FIGURE 62 CUMULATIVE VIEW FOR GAS. ....................................................................................... 66

FIGURE 63 CUMULATIVE VIEW FOR OIL. ......................................................................................... 66

FIGURE 64 CUMULATIVE VIEW FOR WATER. .................................................................................. 67

FIGURE 65 CUMULATIVE VIEW FOR GAS. ....................................................................................... 67

FIGURE 66 DAILY ERROR MPFM AFTER CORRECTION OIL PRODUCTION. ............................... 68

FIGURE 67 MPFM AFTER CORRECTION OF WATER PRODUCTION. ............................................ 68

FIGURE 68 MPFM AFTER CORRECTION OF GAS PRODUCTION. ................................................. 69

FIGURE 69 FPSO OIL PRODUCTION. ................................................................................................ 69

FIGURE 70 FPSO WATER PRODUCTION. ......................................................................................... 70

FIGURE 71 FPSO GAS PRODUCTION. .............................................................................................. 70

FIGURE 72 DAILY ESTIMATION ERROR OF THE OIL PRODUCTION. ............................................ 71

FIGURE 73 DAILY ESTIMATION ERROR OF THE WATER PRODUCTION. ..................................... 72

FIGURE 74 DAILY ESTIMATION ERROR OF THE GAS PRODUCTION. .......................................... 73

FIGURE 75 DAILY ESTIMATION ERROR OF THE RAW OIL PRODUCTION. .................................. 74

FIGURE 76 DAILY ESTIMATION ERROR OF THE RAW OIL PRODUCTION. .................................. 74

FIGURE 77 DAILY ESTIMATION ERROR OF THE RAW OIL PRODUCTION. .................................. 75

FIGURE 78 IFM PERCENTUAL ERROR FOR OIL PRODUCTION. ................................................... 75

FIGURE 79 IFM PERCENTUAL ERROR FOR WATER PRODUCTION. ............................................ 76

FIGURE 80 IFM PERCENTUAL ERROR FOR GAS PRODUCTION. .................................................. 76

FIGURE 81 CUMULATIVE FRECUENCY OF P1 FOR THE WELL ZNA-E0A ..................................... 80

FIGURE 82 CUMULATIVE FRECUENCY OF P2 FOR THE WELL ZNA-E0A ..................................... 80

FIGURE 83 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL ZNA-E0A....................... 81

FIGURE 84 CUMULATIVE FRECUENCY OF P1 FOR THE WELL ZNA-E0D .................................... 81

FIGURE 85 CUMULATIVE FRECUENCY OF P2 FOR THE WELL ZNA-E0D .................................... 82

FIGURE 86 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL ZNA-E0D ...................... 82

FIGURE 87 CUMULATIVE FRECUENCY OF P1 FOR THE WELL ZNA-E0E ..................................... 83

FIGURE 88 CUMULATIVE FRECUENCY OF P2 FOR THE WELL ZNA-E0E ..................................... 83

FIGURE 89 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL ZNA-E0E....................... 84

FIGURE 90 CUMULATIVE FRECUENCY OF P1 FOR THE WELL ZNA-E0H .................................... 84

Page 12: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

12

FIGURE 91 CUMULATIVE FRECUENCY OF P2 FOR THE WELL ZNA-E0H .................................... 85

FIGURE 92 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL ZNA-E0H ..................... 85

FIGURE 93 CUMULATIVE FRECUENCY OF P1 FOR THE WELL ZNA-EA0 ..................................... 86

FIGURE 94 CUMULATIVE FRECUENCY OF P2 FOR THE WELL ZNA-EA0 ..................................... 86

FIGURE 95 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL ZNA-EA0....................... 87

FIGURE 96 CUMULATIVE FRECUENCY OF P1 FOR THE WELL ZNA-EAA .................................... 87

FIGURE 97 CUMULATIVE FRECUENCY OF P2 FOR THE WELL ZNA-EAA .................................... 88

FIGURE 98 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL ZNA-EAA ...................... 88

FIGURE 99 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-F0BA .................................. 89

FIGURE 100 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-F0BA ................................ 89

FIGURE 101 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-F0BA .................. 90

FIGURE 102 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-F0G .................................. 90

FIGURE 103 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-F0BA ................................ 91

FIGURE 104 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-F0BA .................. 91

FIGURE 105 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-FA0 ................................... 92

FIGURE 106 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-FA0 ................................... 92

FIGURE 107 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-FA0 .................... 93

FIGURE 108 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-FAB .................................. 93

FIGURE 109 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-FAB .................................. 94

FIGURE 110 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-FAB .................... 94

FIGURE 111 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-F1C .................................. 95

FIGURE 112 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-F1C .................................. 95

FIGURE 113 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-F1C .................... 96

FIGURE 114 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-FAF .................................. 96

FIGURE 115 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-FAF .................................. 97

FIGURE 116 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-FAF .................... 97

FIGURE 117 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-FAI .................................... 98

FIGURE 118 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-FAI .................................... 98

FIGURE 119 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-FAI...................... 99

FIGURE 120 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-F0E ................................... 99

FIGURE 121 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-F0E ................................. 100

FIGURE 122 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-F0E .................. 100

FIGURE 123 CUMULATIVE FRECUENCY OF P1 FOR THE WELL PRP-F0F ................................. 101

FIGURE 124 CUMULATIVE FRECUENCY OF P2 FOR THE WELL PRP-F0F ................................. 101

FIGURE 125 CUMULATIVE FRECUENCY OF P1 AND P2 FOR THE WELL PRP-F0F ................... 102

FIGURE 126 DEVIATION (P2-P1)/P1 AT WELL ZNA-EA .................................................................. 104

FIGURE 127 DEVIATION (P2-P1)/P2 AT WELL ZNA-EA .................................................................. 104

FIGURE 128 DEVIATION (P2-P1)/WATERCUT1 AT WELL ZNA-EA ................................................ 105

FIGURE 129 DEVIATION (P2-P1)/GOR1 AT WELL ZNA-EA ............................................................ 105

FIGURE 130 DEVIATION (P2-P1)/P1 AT WELL ZNA-ED .................................................................. 105

Page 13: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

13

FIGURE 131 DEVIATION (P2-P1)/P2 AT WELL ZNA-ED .................................................................. 106

FIGURE 132 DEVIATION (P2-P1)/WATERCUT1 AT WELL ZNA-ED................................................ 106

FIGURE 133 DEVIATION (P2-P1)/GOR1 AT WELL ZNA-ED ............................................................ 106

FIGURE 134 DEVIATION (P2-P1)/P1 AT WELL ZNA-EEB ............................................................... 107

FIGURE 135 DEVIATION (P2-P1)/P2 AT WELL ZNA-EEB ............................................................... 107

FIGURE 136 DEVIATION (P2-P1)/WATERCUT1 AT WELL ZNA-EEB ............................................. 107

FIGURE 137 DEVIATION (P2-P1)/GOR1 AT WELL ZNA-EEB .......................................................... 108

FIGURE 138 DEVIATION (P2-P1)/P1 AT WELL ZNA-EH .................................................................. 108

FIGURE 139 DEVIATION (P2-P1)/P2 AT WELL ZNA-EH .................................................................. 108

FIGURE 140 DEVIATION (P2-P1)/WATERCUT1 AT WELL ZNA-EH................................................ 109

FIGURE 141 DEVIATION (P2-P1)/GOR1 AT WELL ZNA-EH ............................................................ 109

FIGURE 142 DEVIATION (P2-P1)/P1 AT WELL ZNA-EA0 ................................................................ 109

FIGURE 143 DEVIATION (P2-P1)/P2 AT WELL ZNA-EA0 ................................................................ 110

FIGURE 144 DEVIATION (P2-P1)/WATERCUT1 AT WELL ZNA-EA0 .............................................. 110

FIGURE 145 DEVIATION (P2-P1)/GOR1 AT WELL ZNA-EA0 .......................................................... 110

FIGURE 146 DEVIATION (P2-P1)/P1 AT WELL ZNA-EAA ............................................................... 111

FIGURE 147 DEVIATION (P2-P1)/P2 AT WELL ZNA-EAA ............................................................... 111

FIGURE 148 DEVIATION (P2-P1)/WATERCUT1 AT WELL ZNA-EAA ............................................. 111

FIGURE 149 DEVIATION (P2-P1)/GOR1 AT WELL ZNA-EAA .......................................................... 112

FIGURE 150 DEVIATION (P2-P1)/P1 AT WELL PRP-F0BA ............................................................. 112

FIGURE 151 DEVIATION (P2-P1)/P2 AT WELL PRP-F0BA ............................................................. 112

FIGURE 152 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-F0BA ........................................... 113

FIGURE 153 DEVIATION (P2-P1)/GOR1 AT WELL PRP-F0BA ........................................................ 113

FIGURE 154 DEVIATION (P2-P1)/P1 AT WELL PRP-F0G ............................................................... 113

FIGURE 155 DEVIATION (P2-P1)/P2 AT WELL PRP-F0G ............................................................... 114

FIGURE 156 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-F0G ............................................. 114

FIGURE 157 DEVIATION (P2-P1)/GOR1 AT WELL PRP-F0G .......................................................... 114

FIGURE 158 DEVIATION (P2-P1)/P1 AT WELL PRP-FA0 ................................................................ 115

FIGURE 159 DEVIATION (P2-P1)/P2 AT WELL PRP-FA0 ................................................................ 115

FIGURE 160 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-FA0 .............................................. 115

FIGURE 161 DEVIATION (P2-P1)/GOR1 AT WELL PRP-FA0 .......................................................... 116

FIGURE 162 DEVIATION (P2-P1)/P1 AT WELL PRP-FAB ............................................................... 116

FIGURE 163 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-FAB ............................................. 116

FIGURE 164 DEVIATION (P2-P1)/GOR1 AT WELL PRP-FAB .......................................................... 117

FIGURE 165 DEVIATION (P2-P1)/P1 AT WELL PRP-FAC ............................................................... 117

FIGURE 166 DEVIATION (P2-P1)/P2 AT WELL PRP-FAC ............................................................... 117

FIGURE 167 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-FAC ............................................. 118

FIGURE 168 DEVIATION (P2-P1)/GOR1 AT WELL PRP-FAC ......................................................... 118

FIGURE 169 DEVIATION (P2-P1)/P1 AT WELL PRP-FAF ................................................................ 118

FIGURE 170 DEVIATION (P2-P1)/P2 AT WELL PRP-FAF ................................................................ 119

Page 14: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

14

FIGURE 171 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-FAF ............................................. 119

FIGURE 172 DEVIATION (P2-P1)/GOR1 AT WELL PRP-FAF .......................................................... 119

FIGURE 173 DEVIATION (P2-P1)/P1 AT WELL PRP-FAI ................................................................. 120

FIGURE 174 DEVIATION (P2-P1)/P2 AT WELL PRP-FAI ................................................................. 120

FIGURE 175 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-FAI ............................................... 120

FIGURE 176 DEVIATION (P2-P1)/GOR1 AT WELL PRP-FAI ........................................................... 121

FIGURE 177 DEVIATION (P2-P1)/P1 AT WELL PRP-F0E ................................................................ 121

FIGURE 178 DEVIATION (P2-P1)/P2 AT WELL PRP-F0E ................................................................ 121

FIGURE 179 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-F0E .............................................. 122

FIGURE 180 DEVIATION (P2-P1)/GOR1 AT WELL PRP-F0E .......................................................... 122

FIGURE 181 DEVIATION (P2-P1)/P1 AT WELL PRP-F0F ................................................................ 123

FIGURE 182 DEVIATION (P2-P1)/P2 AT WELL PRP-F0F ................................................................ 123

FIGURE 183 DEVIATION (P2-P1)/WATERCUT1 AT WELL PRP-F0F .............................................. 123

FIGURE 184 DEVIATION (P2-P1)/GOR1 AT WELL PRP-F0F .......................................................... 124

FIGURE 185 NOISE AND ERRORS IN THE RAW PRODUCTION DATA. ....................................... 125

FIGURE 186 NOISE CLEANING ALLOWS THE GLOBAL BEHAVIOR OF THE CURVE TO BE

IDENTIFIED. ................................................................................................................................ 127

FIGURE 187 PRODUCTION DATA AFTER NOISE CLEANING AND GROUP COMPATIBILITY IN

BETWEEN POINTS. .................................................................................................................... 127

FIGURE 188 FITTED LINE CHOOSEN FOR EACH GROUP OF POINTS. ...................................... 129

FIGURE 189 ORIGINAL POINTS IN THE FINAL GROUPS. RESULT OF A DERIVATIVE

TOLERANCE DTOL=100 AND A GAP TOLERANCE = 10%. .................................................... 129

FIGURE 190 BAR CHART FOR GROUP IDENTIFICATION. ............................................................ 129

FIGURE 191 ORIGINAL PRODUCTION RAW DATA ........................................................................ 130

FIGURE 192 PRODUCTION CURVE SMOOTHED BY THE MOVING AVERAGE ALGORITHM. ... 130

FIGURE 193 PRODUCTION CURVE SMOOTHED BY A STRONGER AVERAGING AND MORE

FLEXIBLE LIMITS FOR CONNECTING GROUPS. .................................................................... 131

FIGURE 194 PRODUCTION CURVE SMOOTHED BY A STRONGER AVERAGING AND MORE

FLEXIBLE LIMITS FOR CONNECTING GROUPS. .................................................................... 131

Page 15: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

15

List of Acronyms

ACA- Acácia (wells)

BHP- Bottom Hole Pressure

BHT- Bottom Hole Temperature

IPM- Integrated Production Management

IFM- Integrated Field Management

FPSO- Floating Production Storage Offloading

GOR- Gas to Oil Ratio

MPFM- Multiphase Flow Meter

P10- Production line 10

P20- Production line 20

P30- Production line 30

P40- Production line 40

P1- Directly measured pressure

P2- Estimated Pressure

PRP- Perpétua (wells)

Q- Total Discharge

Qgas- Gas Discharge

Qoil- Oil Discharge

Qwater- Water Discharge

Rso- Solution Gas-oil Ratio

THP – Tubing Hole Pressure

WHP - Well Head Pressure

WHT- Well Head Temperature

ZNA- Zínia (wells)

Page 16: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

16

Chapter 1. Introduction

The automatic determination of the produced volumes in a subsea level is associated to high ranges

on uncertainty and sereval technical limitations on measurements. Multiphase flow meters are

currently used in the most-advanced facilities as a solution for the problem of the inflow-rate metering

in subsea wells in real time, which is critical for reservoir characterization and well optimization.

Nevertheless, testing, calibrating and post-installation tuning of MPFMs is expensive and can be

problematic. Regular maintenance is often needed and malfunctioning can be constant on these

devices performance. Because of these, drifts, frozen values, data sources failure, out-of-range values

and signal anomalies are often observed on measurements.

It is key that Engineers assume an active role on this systematic measurement process. Attention and

the criticism are needed not only to spot but also to infer on what would be a reasonable solution in a

presence of an existing error.

In the present domain, the Production Allocation Problem regarding the subsea multiphase metering,

some work was previously done in order to minimize the uncertainty from the readings. New

equipment (software) was developed in order to overcome many of the physical metering limitations

and bringing consistency to the data. Nevertheless, these equipments also need maintenance as the

whole system works in a non-stationary process, changing its characteristics over time and as the

reservoir depletes.

Using an investigative strategy, the production data was analyzed in order to build trends that would

be consistent with the physical parameters measured during production. This was done through the

combination of two main different tools: Integrated Field Management (IFM) for the reallocation of

produced volumes and Microsoft Excel for finding production patterns, crosschecking and validating

corrected values. The limitation of IFM relies in the fact that it depends on the accuracy of the inserted

models (PVT, pipelines, etc.). Mainly because of PVT imprecisions it can often provide results based

on biased inputs. Also, if the MPFM values are wrong, that will mean that models are also being built

based on incorrect production volumes. That is why there is the need to change the produced volumes

in the software itself. The IFM works based on two direct measurements: the bottom-hole pressure

and the tubing hole pressure. With these two values it is possible to obtain the pressure drop

associated to the flux and, after, it is possible to run calculations on the total fluids discharge occurring

with the flux – using the Darcy’s law. In the IFM, by knowing the real pressure drop occurring when the

well flows, it is possible to correct the produced volumes that were first measured by the MPFM as the

IFM is also based on density and friction, using the pressure loss on elevation equation.

The challenge was to correct the produced volumes in time so that the estimated pressure drop would

equal the directly measured pressure drop (for which no uncertainties apply). The directly measured

BHP is the key for the solution of this inverse problem: the good calibration of the production

parameters (volumes of oil, volumes of water, volumes of gas, well head pressure and temperature)

Page 17: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

17

relies in the good adjustment in between the directly measured BHP and the estimated BHP (resulting

from the new corrected volumes). This will determine either if the correction is good or not.

The solution for each well was obtained through and iterative process of readjustment in which the

sum of the total produced volumes for oil, water and gas had to match the equivalent amounts

measured with no uncertainty at the FPSO level where it comprises the sum of all the wells of the

field.

In a final stage, it was possible to verify that the correction process apllied improved the volumes

allocation for all the wells. It was also possible to assess the level on uncertainty for each MPFM in

time, in order to forecast the deviated results that are constantly being provided by these devices. The

conclusions obtained during this study also made it possible to spot which PVT models had to be

reviewed and corrected in order to fully assess the reservoirs of the Pazflor field.

This thesis consists in a Monitoring work that pursuits a better forecasting, management and

optimization of the oil field and the resources used on the oil production.

Inflow-rate

metering

Figure 1 Production Reallocation workflow scheme.

Page 18: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

18

Chapter 2. TOTAL'S BLOCK 17

2.1 BLOCK 17 – A World Scale Block

The Block 17 is a world-class oil fields block that covers nearly 4,000 square kilometers (km2) located

from 150 to 270 kilometers off the coast of Angola. This acreage has become the stage for a unique

industrial adventure, with 15 discoveries – developments that have set global benchmarks for the

industry – and a spectacular production outlook. By May 25, 2010, less than a decade after first oil

from the block, its cumulative production had reached 1 billion barrels.

The Block 17 is a large world scale block that represents 1/3 of the total production of the Angolan Oil.

In the following work, although it might seem a system bearing simple characteristics, it is full of

complexities that could only be overcame by sequential synthesis that are made up in order to

approach and understand the processes in a more immediate way.

It began in 1996 with the discovery of Girassol, where a bold field development plan paved the way for

record-setting production of 200,000 barrels per day in 2001. By 2007, the successive tie-backs of the

Jasmim and Rosa fields to the Girassol FPSO, and the start-up of Dalia (one of the largest deepwater

fields ever developed), had boosted the block’s total output to around 500,000 barrels per day. The

Pazflor field came on stream in August 2011 and would quickly ramp up to plateau production. The

development added a further 220,000 barrels per day to cumulative production from Block 17 thanks

to a new world first: the full deployment of a complete subsea artificial lift system with gas/liquids

separation on the seafloor. Growth continued with the 2014 start-up of CLOV, the block’s fourth

production hub, which is contributing with an additional 160,000 barrels per day of oil.

Page 19: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

19

Chapter 3. PAZFLOR

3.1 A Top of the Art Project

The Pazflor project marks the first-ever use of subsea separation technology offshore West Africa, and

it is also a world-first use of subsea separation technology in a development of such scale.

The Pazflor project offshore Angola began producing oil and gas using gas/liquid subsea separation

equipment supplied by FMC Technologies, in September 2011.

Located in the prolific Block 17, some 93 miles (150 km) offshore of Angola, Pazflor produces oil and

gas from 26 subsea wells, supported by 20 water injection and two gas injection wells, drilled in four

separate reservoirs in water depths of 2,000 to 4,000 feet (600 to 1,200 m).

The field covers an area of 238 square miles (600 km2).

A system of flowlines and risers transport the produced fluids to the Floating Production Storage and

Offloading vessel (FPSO) with the processing capacity of 220,000 bopd.

Subsea separation is the key enabling technology making efficient production of this oil possible, as

three of the four Pazflor reservoirs contain very heavy, viscous oil at relatively low reservoir pressures.

The heavy oil will comprise about two-thirds of the liquids produced at Pazflor. Low reservoir pressure

of 2,900 psi (200 bar) meant artificial lift would be needed. Gas lift and multiphase pumping were

assessed but found less efficient than liquid/gas separation with liquid boosting. Subsea separation

allows use of a gas-tolerant hybrid pump to lift the oil and water to the surface, while the gas flows

under its own pressure. Separation results in a more stable flow in the risers.

Separation also simplifies Pazflor’s hydrate prevention strategy. Since the separators operate at a

pressure of 333 psi (23 bar), fluids downstream of each SSU remain outside the hydrate formation

window. Upstream of the SSUs, simple depressurization prevents hydrate formation in the event of an

extended shutdown, without any chemical injection from the surface. The subsea separation system

was instrumental in the development of such an enormous field and available from the first day of

production.

Ultimately, The Pazflor project would not have been feasible without the introduction of Subsea

Separation technology.

3.2 Structure and Equipment

The Pazflor structure was specially conceived with the purpose of meeting all the features needed in

its complex domain. The FPSO has a topside weight of 35,494t.. It is designed with a processing

capacity of 200,000bpd of oil, 150mmcf/d of gas, and a storage capacity of about 1.9 million barrels of

crude.

Page 20: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

20

Facilities are planned for a 20-year design life, and quarters are provided for 220 operation and

maintenance personnel.

Technip's Deep Blue and Deep Pioneer were responsible for engineering, procurement, fabrication

and installation of more than 80km (50 miles) of production and water injection rigid flowlines, flexible

risers and integrated production bundle risers, plus engineering, procurement and fabrication of 60km

(37 miles) of umbilicals.

Acergy was responsible for engineering, procurement, fabrication, and installation of 55km (34 miles)

of water injection, gas injection, and gas export lines, umbilicals, and 20 rigid jumpers, plus installation

of all manifolds, three subsea separation units with umbilicals and FPSO mooring lines.

In the Figure 2, below, a scheme of the subsea producing lines of the Pazflor is presented.

Figure 2 Pazflor producing lines Scheme.

Page 21: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

21

3.3 Measurement of Produced Volumes

3.3.1 The Multiphase Flow Meters (MPFM)

The measurement of the produced volumes is held by Multiphase Flow Meters (MPFM) which are

equipments able to deduce the proportion and flow rate of each fluid phase (oil, water and gas) per

well.

MPFM’s are chosen according to the type of measurement, performance, medium to be applied on

and accuracy requirements. The velocity, pressure, temperature and needs in terms of calibration and

monitoring are also took into account when picking the most adequate type. Another crucial aspect

affecting the performance of these devices regards the ability to overcome the presence of substances

that can be seen as contaminants of the flow, such as injected chemicals, and to be able to read when

in salty and acidic environments. In the Figure 3, below, two examples of MPFM's are present (on the

right) and its location relatively to the FPSO and the umbilicals is also shown (on the left).

The MPFMs are installed in the wellhead of each oil producer well of Pazflor.

For the application of custody transfer measurements of fluid hydrocarbons, positive displacement

meters and turbine meters have been preferred. For gas metering, gas orifice meters and ultrasonic

flow meters are most common. Coriolis meters are in use for liquid measurements, but can also take

gas measurement applications.

For the application of allocation measurements, multiphase flow meters have been adopted, especially

for subsea production systems - as for Pazflor.

A number of factors have instigated the recent rapid uptake of multiphase measurement technology.

In Pazflor, all the oil producer wells are equipped with MPFMs.

3.3.1.1 MPFM Measuring Principles

The MPFM is equipped with a nucleonic sensor that emits gamma-rays through the produced mixture

- from an emissor to a receptor-, the energy that arrives at the receptor allows it to calculate a loss of

energy to me medium that is after associated to a permitivitty value that is possible to link to a certain

amount of hydrocarbons present in the flow.

Figure 3 Example of MPFM's and illustration of its location subsea.

Page 22: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

22

The MPFM works in two different modes. It automatically switches from the Oil Continous mode to the

Water Continuous Mode (or vice-versa). The Water Continuous Mode is activated when the Watercut

is greater than 45% and the measurements are possible measuring the conductivity of the fluid

present in the flow. When working in the Oil Continous Mode, the MPFM estimates de fractions of the

differente phases present in the flow through the measurement of the permitivitty of the fluid.

The Venturi tube is also a fundamental part of the MPFM. According to the Venturi effect, a fluid

passing through smoothly varying constrictions experience changes in velocity and pressure. These

changes can be used to measure the flowrate of the fluid.

As a liquid flows through an orifice, the square of the fluid velocity is directly proportional to the

pressure differential across the orifice and inversely proportional to the specific gravity of the fluid. The

greater the pressure differential, the higher the velocity; the greater the density, the lower the velocity.

The volume flow rate for liquids can be calculated by multiplying the fluid velocity times the flow area.

By taking into account units of measurement, the proportionality relationship previously mentioned,

energy losses due to friction and turbulence, and varying discharge coefficients for various types of

pipelines configurations, the flow rate equation can be written as follows:

(Eq. 1)

In the scheme bellow, Figure 5, it is possible to see a resume of the MPFM Measuring Principles.

Figure 4 Simplified illustration of the Venturi tube.

Figure 5 The MPFM Measuring Principles.

Page 23: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

23

The readings of the MPFMs are stored and daily updated on the PI (Processor that includes all the

data of the FPSO, well by well, second by second) after that it is automatically inserted in the

Production Data Management System (PDMS) and it can be imported either in Flores (in-house

reservoir database currently out of use), Excel or T-More (software database being currently

developed).

Once in Excel, it is possible to plot the curves representing the produced volumes of Oil, Water and

Gas.

These MPFMs offer substantial economic and operating advantages over their phase separating

predecessor. Nevertheless, it is still widely recognized that no single MPFM on the market can meet

all multiphase metering requirements.

In the industry, the way forward and development trends on MPFM can be resumed as the following:

Accuracy improvement for fiscal allocation applications

Low cost systems

Non gamma-ray systems

New technology: Tomography/Imaging

Downhole measurements

Page 24: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

24

Chapter 4. Production Allocation

4.1 The Need and the Goal

The Allocation process is defined by breaking down measures of quantities of extracted hydrocarbons

across the different wells of the field, from one or more reservoirs. In some cases, such as Pazflor, the

well extracts hydrocarbons from more than one geologic formation or reservoir; hence it is useful to

divide the oil field and its well streams by formations or layers. Allocation is commercially rooted in the

need to distribute the costs, revenues and taxes among multiple players collaborating on field

development and production of oil and gas.

There are various incentives for collaborations in the oil industry: one is risk and cost, sharing the

practice by issuing licenses for exploration and production to a partnership of oil companies; another is

the aim of improving production efficiency, producing from multiple layers or multiple oil fields following

the most optimized strategy.

The final goal of this work is to define the most correct production value per well (and per phase of oil,

water and gas), from an extensive set of sensors with inherent uncertainties and showing biased

results.

The understanding of the main trends will lead to better decisions on which wells to either open or

close and where to invest the operating resources.

The objective is to elaborate principles fitting trends that are supported by physical assumptions

regarding the multiphase flow in pipelines, letting the human intervention overcome the effect of drifts

by proposing expressions that eliminate discrepancies and bring consistency to the data.

4.2 The Software - Integrated Production Modelling (IPM) by Petroleum

Experts

'Petroleum Experts’ is a french software company that were one of the pioneers in the direction of

integrating reservoir, production, process and facilities studies. They brought the Integrated Production

Modeling software (IPM) to the market, an outstanding tool that comprises their industry reference

softwares Prosper, GAP, MBAL and IFM.

PROSPER is a well performance, design and optimization program for modelling most types

of well configurations found in the worldwide oil and gas industry today.

GAP is a multiphase oil and gas optimizer tool that models the surface gathering network of

field production systems.

MBAL helps the engineer better define reservoir drive mechanisms and hydrocarbon volumes.

Page 25: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

25

IFM integrates models, data and workflows.

IPM models the complete oil and gas production system including reservoir, wells and the

surface network.

The unique global optimization approach permits the engineer to determine the optimum setting to

maximum production or revenue, taking account of all constraints that are set in the system. These

results can then be used to implement adjustments at the field level to achieve the optimization goals.

4.2.1.1 Integrated Fiel Management (IFM)

The correction of the produced flow rates were held on IFM. This plataform integrates models, data

and workflows, allowing field management tasks to be organized and expedited with maximum

efficiency, from real-time field surveillance tasks to probabilistic forecasting using computer clusters.

The IFM allows engineers to monitor field performance at any time, with continuous real-time

monitoring and surveillance workflows; it automatizes repetitive tasks, allowing engineers to focus on

added-value tasks and it includes the creation and management of multiple scenarios, ensuring the

field optimization and forecasting.

IFM's design is rooted in the belief that sound production system models (i.e. reservoir, well, network

and downstream models) can be used to gain an invaluable understanding of the field's behavior.

Models are a solid basis for field management tasks, such as: field production optimization; well rate

allocation (and consequently a field-wide production allocation); production forecasts; and real-time

production monitoring and surveillance.

Models provide the basic physics that relate the behavior of process variables in reservoirs, wells,

production networks and process facilities. However, a producing field is a dynamic process, and

therefore is constantly changing. Well tests, field measurements, and operational events are field

readings that require proper validation before field management tasks can be undertaken. Figure 6

resumes the Models interation on IFM.

Figure 6 IFM and Models interaction.

Page 26: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

26

ModelCatalogue interacts with IFM, providing it with accurate models at any time. This interaction is

done automatically, and ModelCatalogue acts as the 'access door' to any models IFM may need.

4.2.1.2 Well Rate Allocation Workflow

Traditional methods allocate production according to the available well test data. By continuously

assessing well rates through comparison with calibrated well models, production allocation can be

significantly improved.

4.3 IFM Process - Correction Methodology

The image Figure 8 shows the Well Historical Analysis Workflow (WHAW) used to rebuild the

production history of ZNA-ED well. The top and middle plots represent the liquid rate and the Watercut

respectively as given by the MPFM (in Purple) and by the corrected scenario (in blue), the bottom plot

represents the difference between the directly measured BHP and the estimated using the Prosper

Model and the MPFM flow rates.

The green box represents the maximum admissible difference (+/- 3 bar) in order to take the MPFM

rates estimation as acceptable. When the pressure data is out of the green range it means that the

Watercut and the GOR have to be corrected because the flowing pressure is not coincident with the

real directly measured BHP.

The corrections on the production curves are done by modifying the Watercut and the GOR on the

periods in which the BHP estimated deviates in more than 3 bar from the real BHP (directly measured

through the Down Hole Gauge). These are the only two parameters that are altered in order to obtain

new oil, water and gas rates that are consistent with the directly measured BHP.

Figure 7 Calibration in the Allocation workflow.

Page 27: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

27

The changes are inserted in GOR and Watercut. The impact of these modifications is after possible to

see by selecting the views of the menu on the right.

After running the calculations with the recently modified parameters, it is possible to observe in the

bottom chart of the pannel view (third chart of Figure 8) if the inserted parameters lead us to a

scenario that is consistent with the real measured BHP. If so, the point will be moved to the green

area.

When producing, the well head pressure will decrease as the produced volumes increase and will rise

as they decrease (when the Wells are choked). For the well head temperature the behavior will be the

opposite one: as it will follow the produced volumes and show higher values when it increases and

Figure 8 Rebuilding production history for ZNA-ED well.

Page 28: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

28

lower ones when it decreases. The temperature on the top will be proportional to the producing

volumes and it will only decline if the well is either choked or closed.

Regarding the bottom hole, its pressure is expected to decrease as the produced volumes rise while

its temperature will remain stable around a constant value.

According to these considerations the potential errors are spotted and sent to the Material Balance

team for further revision, analysis and correction of the identified errors before they become “official”.

4.4 IFM Process - The Darcy's Law as the relation in between GOR and

Watercut with Pressure

In fluid dynamics the Darcy's law is a phenomenologically derived constitutive equation that describes

the flow of a fluid through a porous medium. The law was formulated based on the results of

experimnts on the flow of water through beds of sand. Darcy's law is used to describe oil, water and

gas flows through petroleum reservoirs and, in the present study, it is also the expression in which the

IFM calculations are based. When altering the values for the gas flow (GOR) and for the water flow

(Watercut) this expression provides a correspondent result for the pressure drop during

production(from the wellbore to the wellhead).

The Darcy's law presents a simple proportional relationship between the instantaneous discharge rate

through a porous medium, the viscosity of the fluid and the pressure drop over a given distance.

(Eq. 2)

The total discharge, Q (units of volume per time) is equal to the product of the permeability (k, units of

area) of the medium, the cross-sectional area (A) to flow, and the pressure drop (Pb - Pa), all divided

by the dynamic viscosity (in SI units, e.g. Kg/(m.s) or Pa.s), and the length L the pressure drop is

taking place over. The negative sign is needed because the fluids flow from high pressure to low

pressure. Dividing both sides of the equation by the area and using more general notation leads to

(Eq. 3)

where q is the filtration velocity or Darcy flux (discharge per unit area, with units of lenght per time,

m/s) and P is the pressure gradient vector.

For a multiphase flow, an approximation is to use Darcy's law for each phase with permeability

replaced by phase permeability, which is the permeability of the medium multiplied with relative

permeability. This approximation is only valid if the interfaces between the fluids remain static, which is

not true in general, but it is still a reasonable model under steady-stade conditions.

Assuming that the flow of a phase in the presence of another phase can be viewed as single phase

flow through a reduced pore network, it is possible to add the subscript i for each phase to Darcy's law

above written for Darcy flux, and obtain for each phase in the multiphase flow the expression below,

Page 29: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

29

(Eq. 4)

where Ki is the phase permeability for phase i. From this, we can also define relative permeability Kri

for phase i as

(Eq. 5)

where K is the permeability for the porous medium, as in Darcy's law.

The Eq. 4 can be written for , , and . The GOR can be written as the ratio in between

and . It is possible to write the GOR as

(Eq. 5)

And the watercut as

(Eq. 6)

While producing, for each well, the values of the bottom-hole pressure and the tubing-hole pressure

are known. These directly measured values will allow the IFM to calculate the pressure drop (Pa - Pb)

and the pressure gradients . For each well, the permeability of the flowing medium (k), the cross-

sectional area to flow (A), the dynamic viscosity of the fluid ( ) and the length (L) are known and

constant values. That allows the software to calculate possible solutions for the total discharge (Q), for

each Darcy flux per phase and, consequently, the values for the Watercut and the GOR.

Through analysing how the pressure drop (Pb – Pa) varies with the total discharge (Q), it is possible to

infer on how the components of the total discharge, Qoil, Qwater and Qgas, are relatively present in

the flow. If, in a case a), a small variation on the total discharge corresponds to a high variation on the

pressure drop and, in a case b), the same variation on the total discharge corresponds to a smaller

variation on the pressure drop, that means that in the case a) the total discharge has a higher water

component (Qwater) and, in case b), there is a higher gas component (Qgas). Therefore, in case a)

there is more water present in the flow (higher watercut) and, in case b), there is relatively more gas

present in the flow (higher GOR).

THP

BHP

FLUX

Figure 9 Gravity and friction in a flowing well.For the same Qtotal, the pressure drop increases

when there is more water present in the flow.

Page 30: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

30

This is explained by the pressure drop by gravity or vertical elevation equation

(Eq. 7)

where is the pressure drop, is the density of the flowing fluid, g is the acceleration of gravity and

is the vertical elevation.

It is based on the Darcy’s law and on the pressure drop equation that the IFM validates a certain GOR

and Watercut as occurring for each directly measured BHP and THP (pressure drop).

4.5 Correction Criteria for the Reallocation

When applying the corrections in the production flow rates there were different combinations of

Watercut and GOR for a certain period that would be solution to the problem of matching the

estimated pressure with the real pressure. The number of degrees of freedom at this stage would

reflect the level of uncertainty of this Reallocation Problem. For that reason, some correction criteria

had to be set, in order to ensure that the new corrected scenarios would be foundamented and

consistent with the characteristics and constraints of the whole complex system.

The correction was applied according to a conservative approach.

It was only at the last stage, when the data was confronted with the data measured at the FPSO level,

that the new corrected scenarios built under this criteria were sliglty alterred to follow the irrefutable

production curves given by the FPSO for each well.

4.5.1 The Gas-Oil Ratio minimum value

The GOR is the ratio of the volume of gas that is present in the solution to the volume of oil at

standard conditions.

The GOR increases approximately linearly with pressure and is a function of the composition of the oil

and the gas. Heavy oil, as the oil being produced at Pazflor, contains less dissolved gas when

compared to a lighter oil. The GOR increases with pressure, until the bubble point pressure is

reached, after that it remains constant and the oil is said to be undersaturated.

In the IFM, when appliying corrections to the GOR of the produced volumes, it was set that the GOR

corrected values (at the surface) should always be higher tran the value of GOR measured for the

same well at higher depths. As we are producing, it is expected that with the loss of pressure the

Page 31: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

31

volume of gas dissolved in the liquid will increase as a fraction of the oil also changes from the liquid

to the gas phase.

4.5.2 Maximum Parsimony Criteria

The total discharge will be the sum of the oil, water and gas component. The IFM number of solutions

will correspond to the number of combinations of Qoil, Qwater and Qgas that lead to the total

discharge, Q, corresponding to the directly measured pressure drop. For this reason a maximum

parsimony criteria was adopted, stablishing that the correction applied in the GOR and the Watercut

for each well would always be the most conservative one, maintaining the new corrected values of

GOR and Watercut as closest as possible from the original ones.

(Eq. 8) in standard conditions

(Eq. 9) in standard conditions

According to the expressions above, it was a priori expected that the plot of Watercut and the GOR

correction would would present the shape of an exponential function, because in the Watercut

expression the volumes of oil and gas come in the denominator. In Figure 10, an example of how two

different solutions can be obtained for the same well and how they are relatively positioned in terms of

percentual Watercut and GOR corrections.

For every well, the corrected scenario closer to the origin was the scenario validated on IFM.

(Eq. 10)

(Eq. 11)

This was an interative process, that would only stop when no closer solution to the origin would be

possible to find.

Figure 10 Example of two possible corrected scenarios for PRP-FB-A.

%

%

Page 32: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

32

4.5.3 Reservoir Depletion in time

An investigating on the production history of each producing line and well lead us to state a general

trend about the level of certainty of the correction to apply at either a early or late stage of life of an oil

producer well.

For the present stage of production reallocation, it was established that the moment were the Water

Breakthrough was located for each well would be considered as certain. According to the past

acquired experience on production metering and on the MPFM performance, it was known that the

equipment does not show errors in estimation of the flow rates when the Watercut is low. At the

begginning of production, when the MPFM detects a fluid matrix where the hydrocarbons are present

in a higher proportion than water, the MPFM automatically works on the Oil Continuous Mode where it

does the measurements based of the permetivitty of the fluids present in the mixture. Here, the

nucleonic sensor of the MPFM emits gamma-rays through the produced mixture - from an emissor to

a receptor-, the energy that arrives at the receptor allows it to calculate a loss of energy to me medium

that is after associated to a permitivitty value that is possible to link to a certain amount of

hydrocarbons present in the flow. On the other hand, when the Watercut is greater than 45%, the

MPFM automatically switches to the Water Continuous Mode where it does the measurements based

on the conductivity of the fluids present in the mixture. Here, the precision of the MPFM is highly

affected. This problem is related to the fact that it is needed a conductivity of reference to use as input

parameter in order to well calibrate the MPFM. The conductivity is a function of the salinity of the

present water, which has been difficult to assess at the different stages of production. In an advanced

stage of production, after the water breakthrough, as the watercut increases the MPFM estimations

become more deviated: the volumes of water are likely to be considerably underestimated.

For these, a qualitatite strategy was adopted, in order to set in which stage of production would it be

more likely to correct either the GOR, the Watercut or both.

At a early stage of production, before the water breakthrough, if any correction is needed, it should to

be applied on the GOR. After the water breakthrough, when some water starts to show, it corrections

Figure 11 Graphic Illustration of the minimum change criteria applied.

Page 33: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

33

should be logically applied on both. When the Watercut is great than 45%, the corrections are likely to

be as much applied on the GOR as on the Watercut.

4.6 The Correction: MPFM Raw vs MPFM Corrected

4.6.1 Production line: P10

In the Figures 13 and 14 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas). The first figure represents the evolution

corresponding to the liquid phase and the second represents the same evolution in time

corresponding to the gas phase. The objective is to see how different were the allocated flows

(Validated Scenario) from the originally measured flows (Raw MPFM). The exact same approach is

used when analysing wells individually.

According to IFM, the corrections on the flow rates for the P10 line show the following main aspects:

Although the water breakthrough is well allocated, the corrected water rates from October

2012 are approximately 30% higher than registered.

From July 2013 it is possible to see an improvement on metering as the deviation in between

Raw and Corrected values for water rates significantly decreased.

The oil rates deviation are impacted by the water breakthrough. The corrected values are

approximately 7% below the measured ones. Also from July 2013 the values start to converge

again. From November 2014 the deviations increase to again to a maximum value of

approximately 25% for water and approximately 23% for oil in late February 2015.

Figure 12 Graphic Illustration of the criteria for the level of certainty about corrections at a certain stage of the producing time.

Page 34: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

34

The corrected trend for the gas rate shows two different evolution stages. In the first, the

corrected rate for the gas generally matches the measured rates. Here, there is a deviation

around 25% from May 2012 up to October 2012. In the second evolution stage the results

show that the MPFM is generally overestimating the gas rates. From August up to November

2013 it was overestimated in approximately 28%, then it overestimates to a value which is the

double of the corrected rate until May 2014 and from November up to December 2014 it

shows a deviation around 35%. From January 2015 the gas rate values start to converge

again.

0

2000

4000

6000

8000

10000

12000

[Sm3/day]

[Measured time]

MPFM Raw vs MPFM Corrected

Raw MPFMQoil

Raw MPFMQwaater

ValidatedScenario Qoil

ValidatedScenarioQwater

Figure 13 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the line P10.

0

200

400

600

800

1000

1200

1400

[Sm3/day]

[Measured time]

MPFM Raw vs MPFM Corrected

Raw MPFMQgas

ValidatedScenario Qgas

Figure 14 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the line P10.

Page 35: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

35

4.6.1.1 Trend Corrections per Well - Consistency Check

For the wells of the previous producing line (P10) in which strong deviations had to be applied in order

to obtain the corrected rates validated by the IFM, the resulting physical production parameters (WHT,

WHP and BHP) plots are also shown at the end of each section. This way, it is possible to confirm that

the new corrected scenario was consistently built.

4.6.1.1.1 ZNA-E0A

NOTE: For this well, it was not possible to find an adjustable trend from October 2012 until July 2013.

During this period there are no input parameters on IFM that provide a validation of the production

values. Therefore, it is not possible to interpret the results during this period.

Nevertheless, it is possible to confirm that measurements of the MPFM in ZNA-EA are providing good

results for at least approximately 70% of its production period.

Regarding the non-corrected period, it is still possible to observe that the well head temperature

evolution shows an increase from July 2012 until late June 2013- moment when the well was closed-,

while in the BHP pressure trend it is possible to see an increase also from July 2012 which indicates

that less oil is being produced and more water is present on the flow.

In the Figures 15 and 16 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas).

Below, in the Figures 17, 18 and 19 the resulting physical production parameters WHT, WHP and

BHP plots are respectively shown. Confirming a consistent evolution of the 3 parameters for this well

in the new corrected scenario.

Figure 15 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the well ZNA-E0A

Page 36: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

36

Figure 16 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the well ZNA-E0A

Figure 17 Well head temperature evolution in time for the well ZNA-E0A

(ºC)

Page 37: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

37

Figure 18 well head pressure evolution in time for the well ZNA-E0A

Figure 19 Bottom-hole pressure evolution in time for the well ZNA-E0A

(Bar)

(Bar)

Page 38: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

38

Bellow, in the Fugures 20 and 21 it is possible tho see how the changes in the production parameters

lead to an estimated pressure that well captures the trend of the real pressure.

4.6.1.1.2 ZNA-E0E

The measurements of the MPFM in ZNA-EEB are providing good results for oil and water rates in

approximately 85% of its production period – the misreading occurs from middle of November 2014

until late February 2015 and does not overcome the value of 5% in deviation neither for oil nor water.

This occurs short after the water rates overcome the oil rates. During that period, the oil is

overestimated while the water is underestimated by the MPFM.

0

50

100

150

200

250

300

Pressure (Bar)

[Measured time]

Evolution of the pressure at well head

P1

P2

Figure 20 Evolution of the pressure at the well head. P1 – measured pressure. P2 – estimated pressure.

120

130

140

150

160

170

180

190

200

120 140 160 180 200

P2

P1

P1 vs P2

Figure 21 P1 vs P2 – linear regression. P1 – measured pressure. P2 – estimated pressure.

(Bar)

(Bar)

Page 39: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

39

Here, the core problems occur for the gas rates measurements where it misread the gas volumes

during the whole studied period. It is possible to identify three main periods. In the first two the gas

rates are being underestimated by the MPFM where the deviation in between gas measured and

corrected rates decreases in March 2014 from a value of approximately 38% to a deviation value of

approximately 20%; In the final period, the rates are overestimated in approximately 30%. This last

period is coincident with the misreading period on oil and water, also when the watercut reaches the

value of approximately 50%.

The corrected gas rates evolution is confirmed by the well head temperature that shows two different

increasing slopes that are coincident with the two trends followed by the gas rates in the two first

different stages– also the two stages when the proportion of water in the flow increases, first, with a

higher and, after, with the same slope as the oil rates. It is also possible to identify a decreasing trend

in the plots which is coincident with the third evolution stage.

The both WHP and BHP pressure plots have an inflexion point around early February 2014. This

confirms the fact that it is from this date that the proportion of oil present in the flow decreases

significantly while the water production rate increases – as a confirmation of the corrected scenario.

In the Figures 22 and 23 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas). In the Figures 24, 25 and 26 the resulting

physical production parameters WHT, WHP and BHP plots are respectively shown. Confirming a

consistent evolution of the 3 parameters for this well in the new corrected scenario.

0

600

1200

1800

2400

3000

3600

[Sm3/day]

[Measured time]

MPFM Raw vs MPFM Corrected

Raw MPFMQoil

Raw MPFMQwater

ValidatedScenario Qoil

ValidatedScenarioQwater

Figure 22 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the well ZNA-E0E.

Page 40: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

40

Figure 23 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the well ZNA-E0E

Figure 24 Well head temperature evolution in time for the well ZNA-E0E

(ºC)

Page 41: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

41

Figure 25 Well head pressure evolution in time for the well ZNA-E0E

In the Figures 27 and 28 it is possible tho see how the changes in the production parameters lead to

an estimated pressure that well captures the trend of the real pressure.

Figure 26 Bottom-hole pressure evolution in time for the well ZNA-E0E

(Bar)

(Bar)

Page 42: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

42

Figure 27 Evolution of the pressure at the well head. P1 – measured pressure. P2 – estimated pressure.

Figure 28 P1 vs P2 – linear regression. P1 – measured pressure. P2 – estimated pressure.

4.6.1.1.3 ZNA-E0D

According to the IFM, the measurements of the MPFM in ZNA-ED start to misread when the water

breakthrough takes place.

The WBT is well allocated but the water rate that shows after starts to be approximately 80% higher

than the registered one and, on middle of August 2013, it reaches the value of approximately 27% of

deviation. From September 2013, the registered and corrected measurements start matching again,

showing the maximum deviation of approximately 5% for the oil rates and 13% for the water rates.

This improvement on measurements was verified until late October 2014 when the deviations highly

increased again. On December 2014 it reaches values of approximately 85% and 75%– for the water

rates and the oil rates, respectively.

(Bar)

(Bar)

Page 43: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

43

The gas rates are well allocated until October 2014 when the presence of gas in the flow starts to be

overestimated, reading a value that should be in an average approximately 70% lower than the one

registered by the MPFM.

According to the correction applied, the flow has a lower proportion of oil compared to water ever

since the water breakthrough took place. At that moment the watercut overcame a value of

approximately 40% it is when the higher deviations show for both rates.

The well head temperature trend shows an increase from the moment the change in the rates

evolution starts (increasing volumes of water and lower volumes of oil in the flow). In the WHT, WHP

and BHP plots it is possible to see a point of inflexion in the moment when the corrected water rate

reaches its maximum value after a trend of continuous increase- approximately 3550 Sm3/day in

August 2013. The physical parameters are also consistent with the gas rate corrected trend. It is

possible to see on the BHP plot the impact due to the increase on the gas rate from June 2014.

In the Figures 29 and 30 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas).

Figure 29 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the well ZNA-

E0D.

Page 44: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

44

Figure 30 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the well ZNA-E0D

In the Figures 31, 32 and 33 the resulting physical production parameters WHT, WHP and BHP plots

are respectively shown. Confirming a consistent evolution of the 3 parameters for this well in the new

corrected scenario.

Figure 31 Well head pressure evolution in time for the well ZNA-E0D

(Bar)

Page 45: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

45

Figure 32 Well head temperature evolution in time for the well ZNA-E0D

Figure 33 Bottom-hole pressure evolution in time for the well ZNA-E0D

The biplot on Figure 35 presents a considerable dispersion in the first part of the graph due to fact

that, for this was, the DP Venturi was, during some time, out of range. Because of that, during that

period, there is a lack of data regarding the production parameters for this well and it is impossible on

IFM to estimate a better value for the pressure observed while producing.

(Bar)

(ºC)

Page 46: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

46

0

20

40

60

80

100

120

140

160

180

200

Pressure (Bar)

[Measured time]

Evolution of the pressure at well head

P1

P2

Figure 34 Evolution of the pressure at the well head. P1 – measured pressure. P2 – estimated pressure.

120

130

140

150

160

170

180

190

200

120 140 160 180 200

P2

P1

P1 vs P2

Figure 35 P1 vs P2 – linear regression. P1 – measured pressure. P2 – estimated pressure.

(Bar)

(Bar)

Page 47: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

47

4.6.1.1.4 PRP-F0BA

According to the IFM correction, the measurements of the MPFM in PRP-F0BA for the oil and water

rates were providing values with less than 5% of deviation during approximately 76% of the studied

period– until middle April 2014. The gas rates show higher deviations and during a period that

represents more than 80% of the total production period of the well.

For the Liquid phase, the deviations started on September 2013 when the watercut reached the value

of approximately 37%. In April 2014, when the watercut had the average value of approximately 47%,

the previous 5% of deviation in oil and water rates increased to approximately 10% and 8% for water

and oil rates, respectively.

For the gas rates, the most significant deviation occurred from April 2012 until January 2013 when the

gas was underestimated in approximately 40%. From September 2013 it starts to be overestimated in

approximately 55% of the corrected value- this stage is coincident with the deviations on the Liquid

phase readings as well.

The well head temperature shows an increase by gradually higher slopes which confirm the increase

of the water rates from the beginning of the production. The higher slope and the following stabilization

are coherent with the validated evolution trend for both- oil and water- in the flow. The validated trends

on the IFM are also consistent with the WHP and BHP evolution trends. For the WHP and the BHP,

the moment the water and oil corrected rates intersect is coincident with an inflexion point on both

plots. In the Figures 36 and 37 it is presented the evolution in time of the oil produced flow (Qoil),

water produced flow (Qwater) and the gas produced flow (Qgas).

Figure 36 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the well PRP-

F0BA

Page 48: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

48

Figure 37 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the well PRP-F0BA

In the Figures 38, 39 and 40 the resulting physical production parameters WHT, WHP and BHP plots

are respectively shown. Confirming a consistent evolution of the 3 parameters for this well in the new

corrected scenario.

Figure 38 Well head temperature evolution in time for the well PRP-F0BA

(ºC)

Page 49: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

49

Figure 39 Well head pressure evolution in time for the well PRP-F0BA

Figure 40 Bottom-hole pressure in time for the well PRP-F0BA

(Bar)

(Bar)

Page 50: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

50

Bellow, in Figures 41 and 42 it is possible tho see how the changes in the production parameters lead

to an estimated pressure that well captures the trend of the real pressure.

Figure 41 Bottom-hole pressure evolution in time for the well PRP-F0BA

Figure 42 P1 vs P2 – linear regression. P1 – measured pressure. P2 – estimated pressure.

4.6.2 Production line - P20

In the Figures 30 and 31 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas). For better understanding the figures, the

first figure represents the evolution corresponding to the liquid phase and the second the same

evolution in time corresponding to the gas phase. The objective is to see how different were the

allocated flows (Validated Scenario) from the originally measured flows (Raw MPFM). The exact same

approach is used when analysing wells individually.

According to IFM, the corrections on the flow rates for the P20 line present the following main aspects:

The water breakthrough is not well estimated by the MPFM. According to the correction

applied, it happened approximately 2 months after, around late January 2012.

(Bar)

(Bar)

Page 51: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

51

Generally, the deviations mainly occur in two main stages from middle August 2012 until late

August 2013 and, after, from September 2013 until early February 2014. Where the

proportion of oil compared to water in the flow is, respectively, lower and higher than

registered by the MPFM.

Both reach this maximum value after September 2014 when the MPFM measurements start to

overestimate on the oil rates and to underestimate on the water rates.

The gas rates face a correction that is more extended in time. Until September 2013 the

results show that the MPFM was under estimating the gas rates. The trend of gas production

was well captured by the MPFM. Nevertheless, the measured rates would be approximately

10% below the correct gas rate and, from late October 2012 this deviation would significantly

increase to a maximum value of approximately 15% in late August 2013. After that, from

October 2013 the MPFM started to continuously overestimate the gas rate with a deviation

that would reach the value of approximately 25% in February 2015.

In the Figures 43 and 44 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas) for all the wells of the P20 line.

0

2000

4000

6000

8000

10000

12000

14000

16000

[Sm3/day]

[Measured time]

MPFM Raw vs MPFM Corrected

Raw MPFMQoil

Raw MPFMQwater

ValidatedScenario Qoil

ValidatedScenarioQwater

Figure 43 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the line P20.

Page 52: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

52

Figure 44 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the line P20.

4.6.2.1 Trend Corrections per Well - Consistency Check

For the wells of the previous producing line (P20) in which strong deviations had to be applied in order

to obtain the corrected rates validated by the IFM, the resulting plots are shown below.

4.6.2.1.1 PRP-FA0

According to the IFM correction, the measurements of the MPFM in PRP-FA0 for the oil and water

rates were providing values with less than 5% of deviation only during approximately 10% of its

production period. The corrections applied show that the MPFM experienced different measurement

stages that lead the readings to significant deviations either for the oil, water or the gas rates.

The water breakthrough registered on late November 2011 is now allocated on late January 2012.

After this date, the raw and corrected measurements for oil and water rates are coincident for a period

of 5 months. Like in PRP-F0G, it is also from August 2012 when the higher deviations occur, but, in

the case of this well, the oil rates are underestimated while the water and gas rates are overestimated.

For the liquid rates the deviations show values that reach approximately 50%, 20%, respectively for oil

and water rates until July 2013. The validated scenario shows that the water rates overcame the oil

rates around late August 2012 and the proportion of water compared to oil in the flow is significantly

higher than registered. According to the MPFM, this rates intersection happened in December 2012-

approximately 4 months after. When reopened, from September 2014 until the end of the studied

period, the deviations in between raw and corrected values decreased.

The WHT, WHP and BHP for this well are consistent with the high variations pattern followed during

the production. It is possible to observe the significant increase of the water in the flow by analyzing

these physical parameters.

Page 53: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

53

In the Figures 45 and 46 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas) for all the wells of the P20 line.

Figure 45 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the well PRP-

FA0.

Figure 46 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for well PRP-FA0.

In the Figures 47, 48 and 49 the resulting physical production parameters WHT, WHP and BHP plots

are respectively shown. Confirming a consistent evolution of the 3 parameters for this well in the new

correctedscenario.

Page 54: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

54

Figure 47 Well head temperature evolution in time for the well PRP-FA0

Figure 48 Well head pressure evolution in time for the well PRP-FA0

(Bar)

(ºC)

Page 55: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

55

In Figures 50 and 51 it is possible tho see how the changes in the production parameters lead to an

estimated pressure that well captures the trend of the real pressure.

120

130

140

150

160

170

180

190

200

120 140 160 180 200

P2

P1

P1 vs P2

Figure 49 Well head pressure evolution in time for the well PRP-FA0

Figure 50 P1 vs P2

(Bar)

(Bar)

(Bar)

Page 56: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

56

4.6.3 Production line - P30

In the Figures 38 and 39 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas). For better understanding the figures, the

first figure represents the evolution corresponding to the liquid phase and the second the same

evolution in time corresponding to the gas phase. The objective is to see how different were the

allocated flows (Validated Scenario) from the originally measured flows (Raw MPFM). The exact same

approach is used when analysing wells individually.

According to IFM, the corrections on the flow rates for the P30 line present the following main aspects:

The water breakthrough is well estimated by the MPFM.

Generally, the deviations for the oil rates show two different stages. From the beginning of the

production to August 2012 the oil rates are overestimated in approximately 60%; after that

date, the deviation values would only overcome the 10% of deviation from April until

September 2013, from November 2013 until April 2014 and from September 2014 until the

end of the studied period, reaching the maximum values of 30%, 35% and 25% during this

referred periods, respectively.

Regarding the water rates, the maximum deviation occur: first from June 2013 until September

2013, where the presence of water is underestimated in approximately 35%; and after from

September 2013 until September 2014 where it is overestimated in approximately 28% of the

corrected value.

Regarding the gas rates, the IFM shows that the MPFM was not able to capture the gas main

trend. It is possible to observe two different deviation stages, where it is significantly

overestimated and underestimated, being corrected by values that are respectively the half

and the double of the registered ones.

120

130

140

150

160

170

180

190

200

210

Pressure (Bar)

[Measured time]

Evolution of the pressure at well head

P1

P2

Figure 51 Evolution of the pressure at the well head.

Page 57: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

57

The IFM also shows that for this line the corrected water rates overcame the corrected oil

rates around September 2014. According to the MPFM, this intersection happened around

April 2014, 5 months earlier.

In the Figures 52 and 53 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas) for all the wells of the P30 line.

Figure 53 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the line P30.

0

2000

4000

6000

8000

10000

12000

[Sm3/day]

[Measured time]

MPFM Raw vs MPFM Corrected

Raw MPFMQoil

Raw MPFMQwater

ValidatedScenario Qoil

ValidatedScenarioQwater

Figure 52 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the line P30.

Page 58: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

58

4.6.4 Production line - P40

In the Figures 40 and 41 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas). For better understanding the figures, the

first figure represents the evolution corresponding to the liquid phase and the second the same

evolution in time corresponding to the gas phase. The objective is to see how different were the

allocated flows (Validated Scenario) from the originally measured flows (Raw MPFM). The exact same

approach is used when analysing wells individually.

For the Oligocene, the wells ACA805, ACA813 and ACA814 are not possible to correct on the IFM

due to the lack of down hole gauge. Modifications cannot be applied on any of the segments of the

studied period, independently on when the gauge was lost.

For this reason, the validated scenarios for these 3 wells are coincident with the original MPFM raw

scenarios.

Due to the fact that the contribution on oil, water and gas rates of these 3 wells was not possible to be

allocated, building a consistent trend evolution scenario for the P40 line lead to significant deviation

values either for oil, water or gas rates. These deviations possibly show the gaps left by the volume

rates that were not possible to account for. For the gas rates it is possible to observe a constant

underestimation, where the measured rates are lower than the corrected rates in an average value of

approximately 60%. For the liquid rates, the deviation in between raw and corrected values follows the

pattern shown in the image below.

In the Figures 52 and 53 it is presented the evolution in time of the oil produced flow (Qoil), water

produced flow (Qwater) and the gas produced flow (Qgas) for all the wells of the P40 line.

Page 59: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

59

Figure 55 Comparison of the Gas flux for the MPFM raw and the MPFM corrected for the line P40.

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

[Sm3/day]

[Measured time]

MPFM Raw vs MPFM Corrected

Raw MPFMQoil

Raw MPFMQwater

ValidatedScenario

ValidatedScenarioQwater

Figure 54 Comparison of the Water and Oil flux for the MPFM raw and the MPFM corrected for the line P40.

Page 60: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

60

Chapter 5. Method Robustness

In order to infer about the validation of the empirical method applied, it was done multivariate

approaches for simultaneously observing the effect of the different variables.

As presented in the previous section, the analysis of the linear regression resulting of the biplot of the

real pressure versus the estimated pressure allowed to infer about how centered the estimation was

on the variable of control (P1).

A univariate analysis was also performed. Histograms were built in order to see in which classes of

values were the values of P1 and P2 located and evaluate if the distribution of both values had a

similar shape. Throught calculating the plots of the cumulative frecuencies for both variables P1 and

P2 it was possible to see that both cumulative curves would be practically coincident for all the studied

wells.

The Figures from 81 up to 125, in Annex I. show the best P1 and P2 matching for every well of the

Pazflor oil field.

As in a multivariate experiment, varying parameters simultaneously, rather than one at a time, can be

more efficient and can allow the effects between parameters to be observed. A series of plots were

calculated in order to infer on how the deviations in between P1 and P2 would vary with P1 (real

pressure), P2 (estimated pressure), Watercut1 and the GOR1. It was possible to observe in this plots

that the deviation values were independent of the values either of P1, P2, Watercut1 or GOR1. This

prooves the robustness of the method.

The analitycal results of the empirical method applied show it is not biased.

The plots can be consulted in the Annex II, Figures from 126 up to 184 as well as the deviation plots

linked to the Bivariate Analysis and the histograms for the real Pressure (P1) and the estimated

Pressure through IFM (P2).

Page 61: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

61

Chapter 6. Data Reconciliation Results

The FPSO unit comprises de TOPSIDE production level which includes the treatment facilities

responsible for the final stage of the production process, before offloading the oil. The volumes and

rates measurements that take place in the TOPSIDE are more accurate than the measurements held

in the previous levels. This accuracy is well established for the oil volumes measurements but the

confidence level decreases when it is due to either volumes of water or volumes of produced gas. At

the end of each one of the following sections, the measured rates obtained for the different levels –

Subsea Separation Units and Pumps- will be compared to the rates measured on the FPSO. At a final

stage, the MPFM corrected results are superposed to the Topside curves in order to justify its

consistency.

This way the accuracy was assessed and it was possible to infer about the level of calibration not only

on the MPFM’s but also in the SSU’s and Pumps.

The major reference is the Oil curve provided by the measurements on the FPSO oil tanks.

It is important to emphasize that the measurements provided by the Oil thanks are considered to be

100% certain, as it is registered by the fiscal meters before offloading the oil for being sold. The

volumes of water and gas measured at the FPSO level do not have the same level of certainty

associated. As shown in the schemes below, the water and the gas are split in different parts and

porposes in the producing and treatment chain. The accuracy when metering this fractions at the

MPSO is not as high as when metering the main producting product - oil.

For that reason, matching the FPSO Oil production curve with the IFM Oil Corrected Estimation curve

was the main goal during the data reconciliation process that would also take us back to IFM for

readjustments on the corrections previouly applied. Having this curves coincident would also

atomatically mean that the other production curves obtained through the correction on IFM - for water

and gas- would be correctely allocated.

Page 62: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

62

. Figure 56 Above and below, respectively, the resume of the production subsea network and of the topside storage and offloading network.

Page 63: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

63

Chapter 7. Uncertainties and Admissibility

Comparing the oil, water and gas production estimated by the corrected IFM with the FPSO measures

will be fundamental to understand the level of uncertainty of the production allocation strategy and

methodology.

All the devices that register the production rates provide noisy and sometimes intermitent readings. In

the following sections, in order to better capture the production behavior of the combination of wells

there was the need to create a "Curve Repair and Moving Average Algorithm" to clean the noise and

eleminate the errors. The explanation of how this Algorithm was created can be found in Annex III.

Below, we can see for oil, water and gas, respectively, the FPSO production curve, the IFM corrected

estimation, the MPFM Raw estimation curve, and the difference between the FPSO and the IFM

corrected estimation. Respectively in Figures 57, 58 and 59.

-10000

0

10000

20000

30000

40000

50000

[Sm3/day]

[Measured time]

OIL PRODUCTION

OIL-FPSO

OIL-IFM

OIL-RAW MPFM

OIL-FPSO vsOIL-IFM

Figure 57 Comparison of oil production in order of a measured time.

Page 64: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

64

Figure 58 Comparison of water production in order of a measured time.

Figure 59 Comparison of gas production in order of a measured time.

It is on the difference that we should focus, in order to establish sound conclusions about the data and

the study. Because these production curves were full of noise or special events, and we do want to

understand the global behavior without losing the numerical details available, it is interesting to see

the cumulative behavior of both curves (FPSO Production curve and IFM Corrected Estimation curve),

and the difference between them. With the cumulative view, noise will be automatically reduced, and a

Page 65: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

65

global cumulative behavior will be captured if we do accept that random noise will be cumulatively

neglectable with respect to the final results or global trends.

Below, in Figures 60, 61 and 62 we can see the behavior of data in a cumulative view- for oil, water

and gas, respectively.

Figure 61 Cumulative view for water.

-200000

-100000

0

100000

200000

300000

400000

500000

-10 000

0

10 000

20 000

30 000

40 000

50 000

11/01/2011 19/09/2011 27/05/2012 02/02/2013 11/10/2013 19/06/2014 25/02/2015

[Sm3/day]x103 [Sm3/day]

[Measured time]

Cumulative view for oil

OIL-FPSO

OIL-IFM

OIL-FPSO vsOIL-IFM

Figure 60 Cumulative view for oil.

Page 66: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

66

Figure 62 Cumulative view for gas.

Focusing on the differences of the cumulative oil, water and gas production from both curves is

important, but we must acknowledge that this information is not fully consistent, as its meaning will

change along the period of study. As the curve progresses, the meaning of its value gets more

relevant in terms of the total average error achieved in the end, and we cannot see properly the

magnitude of the daily error that we need to observe. Obviously we could observe in the first instance

the initial difference curve from both daily production lines, but unfortunately this curve, because of the

noise, lacks of clarity to be properly evaluated. To overcome this problem the following strategy was

implemented:

First, it is needed to clean the noise of this cumulative difference, so that we can picture it as a smooth

curve. Below we can see the "clean" curve inside the original curve for oil, water and gas -

respectively- Figures 63, 64 and 65.

-200000

-100000

0

100000

200000

300000

400000

[Sm3/day]

[Measured time]

Oil cumulative diference

OIL CUM DIF NONSMOOTH

OIL CUM DIF SMOOTH

Figure 63 Cumulative view for oil.

Page 67: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

67

Figure 64 Cumulative view for water.

Figure 65 Cumulative view for gas.

Page 68: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

68

In order to observe now the “cleaned” differences between The FPSO and the IFM Corrected

Estimation production lines, we can simply calculate a new curve that is in fact the derivative of this

new smoothed cumulative difference curve.

In Figure 66, 67 and 68 it is possible to observe the global evolution of the daily Error, respectively for

oil, water and gas.

-2000

-1000

0

1000

2000

3000

Error [%]

[Measured time]

MPFM after Correction Oil Production Daily Errors

Daily Error (%)

-1000

0

1000

2000

3000

4000

[Sm3/day]

[Measured time]

MPFM AFTER CORRECTION WATER PRODUCTION

Water Absolute Error

Water Absolute Error

Figure 66 Daily error MPFM after correction oil production.

Figure 67 MPFM after correction of water production.

Page 69: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

69

After the average absolute daily error is estimated, it is also needed to understand that the importance

of this error is still dependent of the total daily production values. For that reason, we do want to

understand how much this error represent in terms of percentual deviation for the total oil production.

To perform this we will consider the FPSO production values, but again we will follow the same

strategy, by cleaning the noise of the cumulative FPSO production curve, and in the end calculating

the derivative curve, accepting it as a smooth representation of the occurred production.

The absolute oil, water and gas production measured in the FPSO is shown on Figures 69,70 and 71.

Figure 69 FPSO OIL production.

-400

-200

0

200

400

600

[Sm3/day]

[Measured time]

MPFM AFTER CORRECTION GAS PRODUCTION

Gas Absolute Error

Figure 68 MPFM after correction of gas production.

Page 70: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

70

-1000

4000

9000

14000

19000

24000

29000

34000

[Sm3/day]

[Measured time]

FPSO water production

FPSO Water

0

1000

2000

3000

4000

5000

6000

[Sm3/day]

[Measured time]

FPSO gas production

FPSO Gas

Figure 71 FPSO gas production.

Figure 70 FPSO water production.

Page 71: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

71

Now, it is possible to divide the absolute error by the absolute production and to infer about the

percentual error in the time, for oil, water and gas, respectively, in Figures 72, 73 and 74.

As we can see, there seem to exist 4 relevant periods for the Oil % Daily Estimation Error:

1 – First non-zero production days : Some errors higher than 15 %

2 - Start to 11-02-2012 : Approximately maximum errors of 12 % occurred

3 - 11-02-2012 to 29-04-2104 : Approximately maximum errors of 5 % occurred

4 - 29-04-2104 to End : Approximately maximum errors of 3 % occurred

-10

-5

0

5

10

15

20

Error [%]

[Measured time]

Daily estimation error oil production

Daily Error (%)

Figure 72 Daily estimation error of the oil production.

Page 72: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

72

For the water production, 6 relevant periods for the % Daily Estimation Error can be identified:

1 – First non-zero production days : Some errors lower than 5 %

2 - 10-01-2012 to 06-06-2102 : Errors in between 5% and 12 % occurred

3 - 06-06-2102 to 01-08-2102 : Approximately maximum errors of 5 % occurred

4 - 01-08-2102 to 19-02-2013 : Approximately maximum errors of 9 % occurred

5- 19-02-2013 to 02-07-2013 : Approximately maximum errors of 5 % occurred

6- 02-07-2013 to End : Approximately maximum errors of 14 % occurred

-10

-5

0

5

10

15

20

Error [%]

[Measured time]

Daily estimation error water production

Daily Error (%)

Figure 73 Daily estimation error of the water production.

Page 73: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

73

For the gas production, 5 relevant periods for the % Daily Estimation Error can be identified:

1 – First non-zero production days : Some errors higher than 15 %

2 - 20-08-2011 to 28-08-2013 : Approximately maximum errors of 8 % occurred

3 - 28-08-2013 to 08-06-2014 : Errors in between 8% and 16 % occurred

4 - 08-06-2014 to 22-09-2014 : Approximately maximum errors of 8 % occurred

5- 22-09-2014 to End : Errors lower than 5 %

There are several relevant conclusions:

1 – It seems that the errors have been mitigated along the production period, as the global maximums

have been decreasing consistently;

2 – In global behavior we can say that most of the production estimation has a global error inferior to 5

%;

3 – The oscillating nature of the error in terms of over and under estimating, contributed to the

stabilization of the global accumulated error. However it is not clear the reasons why this behavior is

oscillating, and for that reason, expected global error must be estimated in a conservative manner,

assuming a general percentual error that represents a constant over or under estimation in larger

period of the analysis.

-10

-5

0

5

10

15

20

Error [%]

[Measured time]

Daily estimation error gas production

Daily Error (%)

Figure 74 Daily estimation error of the gas production.

Page 74: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

74

It is also interesting to compare the obtained data with the final non corrected results, using the Raw

MPFM Estimated values, instead the ones updated with IFM.

As before, we can also picture the percentual daily oil, water and gas errors on estimation,

respectively on Figures 75, 76 and 77.

-40

-20

0

20

40

60

80

100

120

Error [%]

[Measured time]

Daily error raw oil production

Daily Error (%)

Figure 75 Daily estimation error of the raw oil production.

0

20

40

60

80

100

120

Error [%]

[Measured time]

Daily Error raw MPFM water production

Daily Error (%)

Figure 76 Daily estimation error of the raw oil production.

Page 75: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

75

In the next plots we will compare the percentual Raw Error in Estimation previously obtained with the

Corrected IFM percentual Error in Estimation. For oil, water and gas, respectively, Figures 78, 79 and

80.

IFM error for Oil

Figure 78 IFM percentual error for oil production.

-60

-40

-20

0

20

40

60

80

100

120

11-18-2010 9-14-2011 7-10-2012 5-6-2013 3-2-2014 12-27-2014 10-23-2015

IFM Corrected % Error

Raw % Error

40

60

80

100

120

140

160

Error [%]

[Measured time]

Daily Error raw MPFM gas production

Daily Error (%)

Figure 77 Daily estimation error of the raw oil production.

Page 76: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

76

The IFM corrections conducted the results to a much acceptable level of error.

-20

0

20

40

60

80

100

120

140

Error (%)

[Measured time]

IFM error for water

IFM CORRECT

RAW

-20

0

20

40

60

80

100

120

140

Error (%)

[Measured time]

IFM error for gas

IFM CORRECT

RAW

Figure 79 IFM percentual error for water production.

Figure 80 IFM percentual error for gas production.

Page 77: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

77

Chapter 8. Conclusions

Several correction cycles and iterations were performed as a consequence of the adopted procedure

of comparing the corrected MPFM production scenarios values with the ones obtained by the Flow

Transmitters available in the FPSO.

The corrected scenarios from the IFM were gradually affected by the difference between each one

and the production curve from the FPSO. At this level, the oil production rate measured at the FPSO

tanks is assumed as a fundamental reference zero uncertainty. As a consequence, it was this final

difference that determined which IFM readjustments were needed at the end of each correction cycle.

This way it was possible to conclude through the calculation of errors in estimation that the produced

volumes of oil, water and gas were being significantly either under or overestimated during the studied

period.

It is important to mention that for matching the corrected MPFM and FPSO production curves it was

necessary to assign GOR values that were below the Rs threshold during several production periods

and for several wells. This can be understood by acknowledging that when facing the need to increase

the total oil rate for a given production volume, it is necessary to decrease simultaneously in the IFM

the amount of water and gas (GOR), obtaining at a next stage the pressure validation by considering

the density and friction in the IFM.

It allows us to conclude that the PVT models need to be reviewed and corrected. The solution gas-oil

ratio is often the most significant component of the PVT correlations. It directly impacts on the oil

formation volume factor (Bo), the oil viscosity and the oil compressibility.

In the Annex IV, Table I, it is possible to consult the Rso table for the Pazflor wells. In green the wells

where the GOR is always above the value of Rso, in red the others.

The IFM corrections conducted the results to a much acceptable level of error in estimation, however it

is important to note the corrected data error in estimation still mimics the oscillating nature of the raw

MPFM error. Even taking in account that IFM corrections are made on individual wells, and that the

final results we are seeing here are the result of the superposition of several wells and lines, this

scaled behavior of the error curve from raw to corrected raises a question: shouldn’t the corrections

drive the results to a more homogenous and constant percentual error? It is not absolutely clear the

answer to this question, but probably the reason for this is the fact that IFM corrections are made over

time segments rather than over individual days. Because of this the convergence of the corrections

may be less precise on individual days, and leading the results to time periods of drag in the direction

of smaller errors, still maintaining the original curve behavior. The relevance of this may be important

because it is likely that a more refined IFM correction will lead to an overall smaller average error.

It is also important to note that the errors of the Raw MPFM data version are relatively high.

Page 78: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

78

In the Integrated Field Management tool, the Models Auditing settings may be not efficiently set up, as

one of its features is to real time estimate production. I would suggest further investigation on how to

optimize the IFM integration in our fields and at the same time, select and define protocols and

procedures to further reduce the production allocation errors.

Chapter 9. Way Forward

In terms of correction strategy and best value estimation, building an artificial neuronal network would

represent a significant adding value approach of the studied problem. Using an artificial neuronal

network to estimate or approximate functions that would simoultaneously depend on the large number

of inputs inherent to this case of study would certainly provide better results in terms of performance,

accuracy and time.

Page 79: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

79

Chapter 10. References

Exploration and Production Internal Files, TOTAL E&P Angola

L.P. Dake, Fundamentals of Reservoir Engineering - Developments in Petroleum Science 8,

Elsevier, Amsterdam (1978)

Prof. Dr. Antonio Costa e Silva, Notes on IST MEP subject of "Oil & Gas"

Hewitt, G.F. (1982) Chapter 2, Handbook of Multiphase Systems (Ed. G. Hetsroni),

Hemisphere Publishing Corporation, New York

Roxar Multiphase Flow Meters, ROXAR

F. Viana, P. Mehdizadeh, Challenges on Multiphase Flow Metering i Heavy Oil Applications,

Society of Petroleum Engineers, Canada (2013)

F. Khan et al, Well-performance Monitoring (WPM): Creating Added VAlue From Raw Data

and Application to the Girassol Deepwater-field Case, Society of Petroleum Engineers, SPE

Economics and Management (2012)

Page 80: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

80

Annexes

Annex I. Cumulative Frequency Graphs

P10

ZNA-E0A

Figure 81 Cumulative frecuency of P1 for the well ZNA-E0A

Figure 82 Cumulative frecuency of P2 for the well ZNA-E0A

Number

of Days

(Bar)

Page 81: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

81

Figure 83 Cumulative frecuency of P1 and P2 for the well ZNA-E0A

ZNA-E0D

Figure 84 Cumulative frecuency of P1 for the well ZNA-E0D

0

200

400

600

800

1000

1200

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 82: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

82

Figure 85 Cumulative frecuency of P2 for the well ZNA-E0D

Figure 86 Cumulative frecuency of P1 and P2 for the well ZNA-E0D

ZNA-E0E

0

200

400

600

800

1000

1200

1400

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 83: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

83

Figure 87 Cumulative frecuency of P1 for the well ZNA-E0E

Figure 88 Cumulative frecuency of P2 for the well ZNA-E0E

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 84: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

84

Figure 89 Cumulative frecuency of P1 and P2 for the well ZNA-E0E

ZNA-E0H

Figure 90 Cumulative frecuency of P1 for the well ZNA-E0H

0

100

200

300

400

500

600

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 85: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

85

Figure 91 Cumulative frecuency of P2 for the well ZNA-E0H

Figure 92 Cumulative frecuency of P1 and P2 for the well ZNA-E0H

ZNA-EA0

0

50

100

150

200

250

300

350

400

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency ofP1Cumulative frequency ofP2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 86: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

86

Figure 93 Cumulative frecuency of P1 for the well ZNA-EA0

Figure 94 Cumulative frecuency of P2 for the well ZNA-EA0

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 87: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

87

Figure 95 Cumulative frecuency of P1 and P2 for the well ZNA-EA0

ZNA-EAA

Figure 96 Cumulative frecuency of P1 for the well ZNA-EAA

0

50

100

150

200

250

300

350

400

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 88: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

88

Figure 97 Cumulative frecuency of P2 for the well ZNA-EAA

Figure 98 Cumulative frecuency of P1 and P2 for the well ZNA-EAA

PRP-F0BA

0

10

20

30

40

50

60

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 89: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

89

Figure 99 Cumulative frecuency of P1 for the well PRP-F0BA

Figure 100 Cumulative frecuency of P2 for the well PRP-F0BA

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 90: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

90

Figure 101 Cumulative frecuency of P1 and P2 for the well PRP-F0BA

P20

PRP-F0G

Figure 102 Cumulative frecuency of P1 for the well PRP-F0G

0

10

20

30

40

50

60

90 100 110 120 130 140 150 160 170

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 91: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

91

Figure 103 Cumulative frecuency of P2 for the well PRP-F0BA

Figure 104 Cumulative frecuency of P1 and P2 for the well PRP-F0BA

PRP-FA0

0

200

400

600

800

1000

1200

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 92: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

92

Figure 105 Cumulative frecuency of P1 for the well PRP-FA0

Figure 106 Cumulative frecuency of P2 for the well PRP-FA0

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 93: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

93

Figure 107 Cumulative frecuency of P1 and P2 for the well PRP-FA0

PRP-FAB

Figure 108 Cumulative frecuency of P1 for the well PRP-FAB

0

200

400

600

800

1000

1200

1400

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 94: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

94

Figure 109 Cumulative frecuency of P2 for the well PRP-FAB

Figure 110 Cumulative frecuency of P1 and P2 for the well PRP-FAB

PRP-F1C

0

100

200

300

400

500

600

700

800

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 95: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

95

Figure 111 Cumulative frecuency of P1 for the well PRP-F1C

Figure 112 Cumulative frecuency of P2 for the well PRP-F1C

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 96: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

96

Figure 113 Cumulative frecuency of P1 and P2 for the well PRP-F1C

PRP-FAF

Figure 114 Cumulative frecuency of P1 for the well PRP-FAF

0

200

400

600

800

1000

1200

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 97: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

97

Figure 115 Cumulative frecuency of P2 for the well PRP-FAF

Figure 116 Cumulative frecuency of P1 and P2 for the well PRP-FAF

PRP-FAI

0

100

200

300

400

500

600

700

800

900

1000

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 98: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

98

Figure 117 Cumulative frecuency of P1 for the well PRP-FAI

Figure 118 Cumulative frecuency of P2 for the well PRP-FAI

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 99: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

99

Figure 119 Cumulative frecuency of P1 and P2 for the well PRP-FAI

P30

PRP-F0E

Figure 120 Cumulative frecuency of P1 for the well PRP-F0E

0

100

200

300

400

500

600

700

800

900

1000

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 100: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

100

Figure 121 Cumulative frecuency of P2 for the well PRP-F0E

Figure 122 Cumulative frecuency of P1 and P2 for the well PRP-F0E

PRP-F0F

0

200

400

600

800

1000

1200

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 101: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

101

Figure 123 Cumulative frecuency of P1 for the well PRP-F0F

Figure 124 Cumulative frecuency of P2 for the well PRP-F0F

Number

of Days

Number

of Days

(Bar)

(Bar)

Page 102: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

102

Figure 125 Cumulative frecuency of P1 and P2 for the well PRP-F0F

0

200

400

600

800

1000

1200

1400

90 110 130 150 170 190

Cumulative frequency P1 and P2

Cumulative frequency of P1

Cumulative frequency of P2

Number

of Days

(Bar)

Page 103: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

103

Page 104: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

104

Annex II. Deviation Graphs

P10

ZNA-EA

Figure 126 Deviation (P2-P1)/P1 at well ZNA-EA

Figure 127 Deviation (P2-P1)/P2 at well ZNA-EA

Page 105: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

105

Figure 128 Deviation (P2-P1)/Watercut1 at well ZNA-EA

Figure 129 Deviation (P2-P1)/GOR1 at well ZNA-EA

ZNA-ED

Figure 130 Deviation (P2-P1)/P1 at well ZNA-ED

Page 106: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

106

Figure 131 Deviation (P2-P1)/P2 at well ZNA-ED

Figure 132 Deviation (P2-P1)/Watercut1 at well ZNA-ED

Figure 133 Deviation (P2-P1)/GOR1 at well ZNA-ED

Page 107: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

107

ZNA-EEB

Figure 134 Deviation (P2-P1)/P1 at well ZNA-EEB

Figure 135 Deviation (P2-P1)/P2 at well ZNA-EEB

Figure 136 Deviation (P2-P1)/Watercut1 at well ZNA-EEB

Page 108: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

108

Figure 137 Deviation (P2-P1)/GOR1 at well ZNA-EEB

ZNA-EH

Figure 138 Deviation (P2-P1)/P1 at well ZNA-EH

Figure 139 Deviation (P2-P1)/P2 at well ZNA-EH

Page 109: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

109

Figure 140 Deviation (P2-P1)/Watercut1 at well ZNA-EH

Figure 141 Deviation (P2-P1)/GOR1 at well ZNA-EH

ZNA-EA0

Figure 142 Deviation (P2-P1)/P1 at well ZNA-EA0

Page 110: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

110

Figure 143 Deviation (P2-P1)/P2 at well ZNA-EA0

Figure 144 Deviation (P2-P1)/Watercut1 at well ZNA-EA0

Figure 145 Deviation (P2-P1)/GOR1 at well ZNA-EA0

Page 111: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

111

ZNA-EAA

Figure 146 Deviation (P2-P1)/P1 at well ZNA-EAA

Figure 147 Deviation (P2-P1)/P2 at well ZNA-EAA

Figure 148 Deviation (P2-P1)/Watercut1 at well ZNA-EAA

Page 112: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

112

Figure 149 Deviation (P2-P1)/GOR1 at well ZNA-EAA

PRP-F0BA

Figure 150 Deviation (P2-P1)/P1 at well PRP-F0BA

Figure 151 Deviation (P2-P1)/P2 at well PRP-F0BA

Page 113: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

113

Figure 152 Deviation (P2-P1)/Watercut1 at well PRP-F0BA

Figure 153 Deviation (P2-P1)/GOR1 at well PRP-F0BA

P20

PRP-F0G

Figure 154 Deviation (P2-P1)/P1 at well PRP-F0G

Page 114: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

114

Figure 155 Deviation (P2-P1)/P2 at well PRP-F0G

Figure 156 Deviation (P2-P1)/Watercut1 at well PRP-F0G

Figure 157 Deviation (P2-P1)/GOR1 at well PRP-F0G

Page 115: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

115

PRP-FA0

Figure 158 Deviation (P2-P1)/P1 at well PRP-FA0

Figure 159 Deviation (P2-P1)/P2 at well PRP-FA0

Figure 160 Deviation (P2-P1)/Watercut1 at well PRP-FA0

Page 116: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

116

Figure 161 Deviation (P2-P1)/GOR1 at well PRP-FA0

PRP-FAB

Figure 162 Deviation (P2-P1)/P1 at well PRP-FAB

Figure 163 Deviation (P2-P1)/Watercut1 at well PRP-FAB

Page 117: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

117

Figure 164 Deviation (P2-P1)/GOR1 at well PRP-FAB

PRP-FAC

Figure 165 Deviation (P2-P1)/P1 at well PRP-FAC

Figure 166 Deviation (P2-P1)/P2 at well PRP-FAC

Page 118: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

118

Figure 167 Deviation (P2-P1)/Watercut1 at well PRP-FAC

Figure 168 Deviation (P2-P1)/GOR1 at well PRP-FAC

PRP-FAF

Figure 169 Deviation (P2-P1)/P1 at well PRP-FAF

Page 119: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

119

Figure 170 Deviation (P2-P1)/P2 at well PRP-FAF

Figure 171 Deviation (P2-P1)/Watercut1 at well PRP-FAF

Figure 172 Deviation (P2-P1)/GOR1 at well PRP-FAF

Page 120: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

120

PRP-FAI

Figure 173 Deviation (P2-P1)/P1 at well PRP-FAI

Figure 174 Deviation (P2-P1)/P2 at well PRP-FAI

Figure 175 Deviation (P2-P1)/Watercut1 at well PRP-FAI

Page 121: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

121

Figure 176 Deviation (P2-P1)/GOR1 at well PRP-FAI

P30

PRP-F0E

Figure 177 Deviation (P2-P1)/P1 at well PRP-F0E

Figure 178 Deviation (P2-P1)/P2 at well PRP-F0E

Page 122: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

122

Figure 179 Deviation (P2-P1)/Watercut1 at well PRP-F0E

Figure 180 Deviation (P2-P1)/GOR1 at well PRP-F0E

Page 123: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

123

PRP-F0F

Figure 181 Deviation (P2-P1)/P1 at well PRP-F0F

Figure 182 Deviation (P2-P1)/P2 at well PRP-F0F

Figure 183 Deviation (P2-P1)/Watercut1 at well PRP-F0F

Page 124: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

124

Figure 184 Deviation (P2-P1)/GOR1 at well PRP-F0F

Page 125: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

125

Annex III. Curve Repair + Moving Averages Algorithm

Summary

For certain quantity measurements, discrete values are obtained in a time ordered fashion. These

series of values, may represent non-random behaviors, and may contain one or more continuous sets

of points, where it is expected a continuous curve and derivative, that can be grouped or separated by

specific characteristics in the beginning or end, as abrupt average or derivative changes.

However, for certain measurement processes, several aspects can decrease the quality of the

obtained points, introducing random errors or noise in the curves, bringing difficulties to read the

information in a clear graphic representation. It will also create a high amount of obstacles to

automatically process the data. These errors can be a result of failures or bad calibration of the

reading entities, but there is an important assumption that most of the time, the behavior is well caught

by the obtained points, and that there may only be necessary to capture clearly the major trends of the

data.

To overcome the problems referenced as noise, wrong readings or simply unwanted data fluctuation,

a simple flexible algorithm was developed to perform data repair and trend identification.

This algorithm has 3 fundamental steps. Point cleaning, group compatibility, and curve smoothing.

The Algorithm

The developed method is based on the assumption that there is no unique correct solution to the

problem, and for that reason, there are a set of parameters that highly influence the behavior of the

algorithm and that can be used to make trials with the data, until the cleaning and classification is

acceptable in the user’s perspective.

To better understand the steps of the method, a simple example is referenced.

Figure 185 Noise and errors in the raw production data.

0

5000

10000

15000

20000

25000

1

45

8

9

13

3

17

7

22

1

26

5

30

9

35

3

39

7

44

1

48

5

52

9

57

3

61

7

66

1

70

5

74

9

79

3

83

7

88

1

92

5

96

9

10

13

1

05

7

11

01

1

14

5

11

89

1

23

3

12

77

Page 126: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

126

In the previous picture, there is clearly one or more well identified trends, and a continuous set of

points introducing noise and errors. One can see easily the global behavior of the curve, but that

behavior is not immediately visible when processing the raw available data.

Step1 – Noise cleaning

The first step has the goal to eliminate points that represent high magnitude changes. High magnitude

changes can be the result of 2 factors:

a) Errors in the measurements;

b) Abrupt changes in the behavior.

There are 2 problems associated with the elimination of these points, but we will see later that it will

not affect negatively the goal of the algorithm:

a) Erratic points may not occur isolated. There may be several “wrong points” in a sequence.

Of course, if this sequence is too long, it will be considered a trend;

b) When the curve changes realistically, eliminating these points may represent a reduction

of the boundary points of a given well identified trend.

The fundamental decisions in this process of elimination are:

a) How many points do we eliminate?

b) What conditions do we use to decide the elimination?

Clearly there is no unique way to perform this step, and the actual implemented algorithm in VBA and

Excel is not meant to solve robustly all types of problem that may follow inside this category of

situations. However a decision was made to use a 2 stage elimination process.

The 2 stage noise cleaning is performed in the following manner:

a) Choose a Derivative Limit D1 as an input parameter. This parameter must be tough in

such a way that it will be smaller than the derivatives near the vicinity of the points where

we can clearly see abrupt changes, but will be higher than the derivatives that occur

naturally around along the small fluctuations of the curve;

b) The algorithm will now calculate for each point a local derivative magnitude. This is simply

done by measuring the absolute left and right derivatives for each point and averaging

both;

c) All points that will have a derivative magnitude superior to D1 will be eliminated. The

elimination does not create any implicit point in the eliminated position. It simply assumes

that the given discrete position has no point value. Points that will appear isolated and

surrounded by eliminated points will also be eliminated;

Page 127: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

127

d) Choose a second Derivative Limit D2 as an input parameter. We will repeat the process,

especially because normally, for the given real situations being solved, the erratic points

usually appear not completely isolated;

e) Recalculate derivatives as in b);

f) Eliminate again points, this time, where the magnitude of the derivative is greater than D2.

Eliminate again points that appear isolated and surrounded by eliminated points.

In this example, using a magnitude of 900 for both D1 and D2 we obtained the following points:

Figure 186 Noise cleaning allows the global behavior of the curve to be identified.

From these points we already have an implicit set of groups. We can see each group as a continuous

sequence of points without interruption. An interruption occurs when one or more point is missing

because of elimination. A clear picture of the groups and interruptions can be seen in the zoom below

in the X axis zone between 500 and 600.

Figure 187 Production data after noise cleaning and group compatibility in between points.

0

5000

10000

15000

20000

25000

0 200 400 600 800 1000 1200 1400

0

5000

10000

15000

20000

25000

500 520 540 560 580 600

Page 128: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

128

Step 2 – Group compatibility

In this stage, we want to reorganize the existing groups in bigger ones, based on the fact that these

isolated groups may in fact represent the same trend. To achieve this regrouping we could opt for

several simple or complex strategies. The actual implementation opted for a very simple solution

described below:

a) For each group calculate the Line Equation parameters A and B (where y(x)=A.x+B) that

best fits the points contained in the group. This best fit is achieved by calculating A and B

to minimize the sum of the squares of the distances between each point and the line;

b) Define a derivative tolerance DTol;

c) Define a gap tolerance GTol as a percentage of the maximum occurred magnitude in the

whole data;

d) For each possible connection between 2 adjacent groups calculate the derivative

difference of the fitted lines between the line of group n and line of group n+1. Define a

condition 1 that is true if this difference is smaller than the defined DTol parameter;

e) For each possible connection between 2 adjacent groups calculate the vertical distance

between 2 points projected in a vertical line, located in the x coordinate that corresponds

to the average between the last x coordinate of the group n and the first x coordinate of

the group n+1. This projection is made by extending the fitted line of each group onto the

referred vertical line. Define a condition 2 that is true is this distance is smaller than the

defined GTol parameter;

f) If condition 1 and condition 2 are met simultaneously, then the 2 groups being tested are

connected and the process continues until no connections can be made.

Below we can see globally and in more detail the points represented not on the original position, but

over each fitted line for each group.

Page 129: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

129

Figure 188 Fitted line choosen for each group of points.

Figure 189 Original points in the final groups. Result of a Derivative tolerance DTol=100 and a Gap tolerance =

10%.

Based on the final connections made, we can represent the original points in the final groups. In this

case it was used a Derivative tolerance DTol=100 and a Gap tolerance = 10%:

Figure 190 Bar chart for group identification.

0

5000

10000

15000

20000

25000

0 200 400 600 800 1000 1200 1400

0

5000

10000

15000

20000

25000

800 850 900 950 1000 1050 1100 1150 1200

0

5000

10000

15000

20000

25000

0 200 400 600 800 1000 1200 1400

Page 130: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

130

To better understand the start and end of each group, a bar chart is pictured below to identify the

groups from 1 to 7.

Figure 191 Original Production Raw Data

Step 3 – Moving average

In this stage, we want to smooth the data, but only inside each identified already cleaned trend group.

To achieve this goal we will define the number of points with the parameter NP to be used for the

moving average calculation. This number must be odd. In this case we used the number of 21.

The algorithm introduces a special restraint to this number of points. In the case that the count of

points in the group is smaller than the double of the NP, than NP must be smaller than half of the

points of the group. With this, we will avoid unnecessary flattening of points in small groups.

Below we can picture the original curve and the curve points, already smoothed by the moving

average algorithm inside each group in 2 versions. The second version used a stronger averaging and

more flexible limits for connecting groups.

Original:

Figure 192 Production curve smoothed by the moving average algorithm.

0

1

2

3

4

5

6

7

1

42

8

3

12

4

16

5

20

6

24

7

28

8

32

9

37

0

41

1

45

2

49

3

53

4

57

5

61

6

65

7

69

8

73

9

78

0

82

1

86

2

90

3

94

4

98

5

10

26

1

06

7

11

08

1

14

9

11

90

1

23

1

12

72

0

5000

10000

15000

20000

25000

1

44

8

7

13

0

17

3

21

6

25

9

30

2

34

5

38

8

43

1

47

4

51

7

56

0

60

3

64

6

68

9

73

2

77

5

81

8

86

1

90

4

94

7

99

0

10

33

1

07

6

11

19

1

16

2

12

05

1

24

8

12

91

Page 131: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

131

Final result version 1:

Figure 193 Production curve smoothed by a stronger averaging and more flexible limits for connecting groups.

Final result version 2:

Figure 194 Production curve smoothed by a stronger averaging and more flexible limits for connecting groups.

Many other solutions could be achieved by changing the algorithm parameters. There is no

assumption that one is better than the other. Only human interpretation can decide if the obtained

solution fits the desired purpose.

0

5000

10000

15000

20000

25000

0 200 400 600 800 1000 1200 1400

0

5000

10000

15000

20000

25000

0 200 400 600 800 1000 1200 1400

Page 132: Production Reallocation of the Pazflor Oil Field...To all the colleagues I’ve always wanted to be around of. To Rafa, that was always there. And, finnally, to Carlos who was heaven

132

Annex IV. Rs Table for the Pazflor wells

Table I. Rs table for the Pazflor wells. In green the wells where the GOR is always above the value of

Rs, in red the others