INJECTION IN ONSHORE & OFFSHORE ENVIRONMENTS
A THESIS SUBMITTED TO THE DEPARTMENT OF
ENERGY RESOURCES ENGINEERING
OF STANFORD UNIVERSITY
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF MASTER OF SCIENCE
Sel¸cuk Fidan
September 2011
c Copyright by Sel¸cuk Fidan 2011
All Rights Reserved
ii
I certify that I have read this thesis and that in my opinion it is fully
adequate, in scope and in quality, as partial fulﬁllment of the degree of
Master of Science in Energy Resources Engineering.
Prof. Anthony R. Kovscek
(Principal Adviser)
I certify that I have read this thesis and that in my opinion it is fully
adequate, in scope and in quality, as partial fulﬁllment of the degree of
Master of Science in Energy Resources Engineering.
Dr. Louis Castanier
iii
iv
Abstract
In the oil industry, the problem of wellbore heat loss during hot ﬂuid injection is clas
sical. Even today the topic is important for practical application of steam injection.
Most thermal reservoir simulators today do not yet take into account heat losses and
pressure drops along the wellbore. Neglecting these items may be acceptable for shal
low reservoirs. For deeper injection wells and injection wells in oﬀshore environments,
however, wellbore heat loss is often signiﬁcant.
Accurate predictions of heat loss, temperature distributions and pressure proﬁle
are essential for modeling steam injection wells. The main goal of this study is to
investigate heat losses along the wellbore during steam injection in both onshore and
oﬀshore environments. Steam quality, steam temperature, steam pressure, and heat
loss values with and without insulation are calculated. In the literature, it is shown
that the Fontanilla and Aziz model [20] yields results in good agreement with ﬁeld
data. The Fontanilla and Aziz approach is used in this study, with an improvement in
the application of twophase ﬂow correlations and the determination of several input
parameters.
The equations describing mass and heat ﬂow are solved in discretized wellbore
framework. Steam properties are incorporated directly. Several twophase ﬂow corre
lations for injection tubing, are used and results are compared. The calculated steam
temperature and steam pressure agree well with the ﬁeld data using the Beggs and
Brill model [13, 14]. Six insulation materials are examined: 1) black aerogel, 2) white
v
aerogel, 3) ﬁberglass, 4) carbon ﬁber, 5) thermolastic insulation and 6) calcium sili
cate. Aerogel insulations present the opportunity to create a superinsulated tubing
that overcomes many limitations of current steam injectors. A Matlab Graphical User
Interface (GUI) is developed, that enables other users to change the input parameters
and visualize the results without going into the details of the calculations.
To our knowledge, no one has predicted the result of noncondensible gas addition
on steam injectors. In this work, a novel approach is introduced for adding non
condensible gas to steam to increase the injection pressure without increasing the
steam temperature. Additional partial pressure is obtained by adding N
2
to the
system. Steam quality, steam temperature, steam pressure, and heat loss calculation
are conducted for steam injection with noncondensible gas (N
2
). Compared with
the case of just steam injection, the steam temperature values are smaller, so are the
amounts of heat loss.
vi
Acknowledgments
First and foremost, I would like to express my sincere gratitude to my adviser Prof.
Anthony R. Kovscek for his time, support, guidance and his unbelievable patience.
Without his support I would not be able to ﬁnish this work, Thank you Tony!
My thanks also go to Dr. Louis Castanier with whom I had helpful discussions
in the early and late stage of this project. I would also like to thank to visiting
Prof. Jan Dirk Jensen to teach us ”Design and analysis of production systems for oil
and gas reservoirs” to understand multiphase ﬂow concept better. My special thanks
go to Prof. Khalid Aziz to provide one of his Master students Fontanilla’s thesis.
I would like to thank both to Turkish National Petroleum Cooperation (T.P.A.O)
and SUPRIA Aﬃliates for their support. It was one of the biggest dream I had
since second year of undergraduate to come to Stanford and study there. It came
true, this success was not only one person’ success it was the success of the several
people in my life and I will brieﬂy talk about those people here. My mother is the
highest priority person in my life not only she raised us with devoting her entire life to
her children but also lack of opportunity she had not to allowed to get educated, she
dedicated herself to her children to get all of them (4 sons) educated and she achieved
this, thank you ’Anne’ (means mom in Turkish). My adviser in Turkey Prof. Dr.
Mustafa Onur who has great impact on my life in terms of his knowledge, support
and trust on me. He is a unique person both in Academic world and personal world.
He always gave his students courage to excel their skills and motivated them to work
vii
hard, because of him I am here, many thanks to him. My adviser Dr. Mehmet Parlar
at Schlumberger during my internship last summer, he gave me an opportunity to
work with him and learn from his experiences, many thanks to him and as well as
my supervisor Dr. Rajesh Chanpura. Now my friends ; Elnur Aliyev, Amar Alshehri,
Mehrdad Honarkhah, Rustem Zaydullin, Alireza Iranshahr, Obi Isebor: I enjoyed
studying with you guys during my stay at Stanford.
It was my fortune to be one of the member of the such a great team SUPRIA
that I have learned a lot and found my women of dreams and got married. My wife,
Wenjuan Lin, is the softest part of my heart and inspiration of my life, and more.
She always supported me, not only several days I stayed at school she never became
angry, but also she showed her love to me every seconds of our life. Because of her
love, I enable to ﬁnish this work! Thank you, CANIM! And my little baby daughter
Su Lin Fidan who brought fun, joy, and energy to our life, thank you little angel;).
viii
IF
If you can keep your head when all about you
Are losing theirs and blaming it on you;
If you can trust yourself when all men doubt you,
But make allowance for their doubting too;
If you can wait and not be tired by waiting,
Or, being lied about, don’t deal in lies,
Or, being hated, don’t give way to hating,
And yet don’t look too good, nor talk too wise;
If you can dream  and not make dreams your master;
If you can think  and not make thoughts your aim;
If you can meet with triumph and disaster
And treat those two imposters just the same;
If you can bear to hear the truth you’ve spoken
Twisted by knaves to make a trap for fools,
Or watch the things you gave your life to broken,
And stoop and build ’em up with wornout tools;
If you can make one heap of all your winnings
And risk it on one turn of pitchandtoss,
And lose, and start again at your beginnings
And never breath a word about your loss;
If you can force your heart and nerve and sinew
To serve your turn long after they are gone,
And so hold on when there is nothing in you
Except the Will which says to them: ”Hold on”;
If you can talk with crowds and keep your virtue,
Or walk with kings  nor lose the common touch;
If neither foes nor loving friends can hurt you;
If all men count with you, but none too much;
If you can ﬁll the unforgiving minute
With sixty seconds’ worth of distance run 
Yours is the Earth and everything that’s in it,
And  which is more  you’ll be a Man my son!
Rudyard Kipling
ix
Dedicated to my father H¨ usn¨ u Fidan (R.I.P)
x
Contents
Abstract v
Acknowledgments vii
1 Introduction 1
1.1 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Literature Review 6
2.1 Emeraude Vapeur : A Steam Pilot in an Oﬀshore Environment . . . . 6
2.2 Marlin Failure Analysis and Redesign . . . . . . . . . . . . . . . . . . 9
2.3 Heat Transmission Mechanisms and Discussion from Authors . . . . . 11
2.3.1 Heat Transmission Mechanisms . . . . . . . . . . . . . . . . . 11
2.3.1.1 Heat Transfer by Conduction . . . . . . . . . . . . . 11
2.3.1.2 Heat Transfer by Convection . . . . . . . . . . . . . 12
2.3.1.3 Heat Transfer by Radiation . . . . . . . . . . . . . . 12
2.3.2 Heat Transmission Discussion from Authors . . . . . . . . . . 13
3 Model Formulation 15
3.1 Heat Loss Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.1.1 Heat Loss from Surface Lines . . . . . . . . . . . . . . . . . . 17
3.1.1.1 With/without Insulation . . . . . . . . . . . . . . . . 17
xi
3.1.2 Heat Loss from Sea Level to Sea Floor . . . . . . . . . . . . . 19
3.1.2.1 With/without Insulation . . . . . . . . . . . . . . . . 19
3.1.3 Heat Loss from Sea Floor to Reservoir . . . . . . . . . . . . . 22
3.1.3.1 With/without Insulation . . . . . . . . . . . . . . . . 22
3.2 Steam Phase behavior calculations . . . . . . . . . . . . . . . . . . . 26
3.3 Two Phase Flow Correlations . . . . . . . . . . . . . . . . . . . . . . 27
3.3.1 Modiﬁed Beggs and Brill Model . . . . . . . . . . . . . . . . . 32
3.3.1.1 FlowPattern Determination . . . . . . . . . . . . . . 33
3.3.1.2 Hydrostatic Pressure Diﬀerence . . . . . . . . . . . . 36
3.3.1.3 Frictional Pressure Loss . . . . . . . . . . . . . . . . 38
3.3.2 Aziz, Govier and Fogarasi Model . . . . . . . . . . . . . . . . 39
3.3.2.1 Flow Pattern Determination . . . . . . . . . . . . . . 39
3.3.2.2 Modiﬁcations . . . . . . . . . . . . . . . . . . . . . . 43
4 Eﬀect of NonCondensable Gas (N
2
) 45
5 Graphical User Interface (GUI) 49
6 Results and Comparisons 53
6.1 Examples for heat loss calculation . . . . . . . . . . . . . . . . . . . . 53
6.1.1 Example 10.1 from Prats . . . . . . . . . . . . . . . . . . . . . 54
6.1.2 Example for Oﬀshore . . . . . . . . . . . . . . . . . . . . . . . 59
6.1.3 Example 10.2 from Prats . . . . . . . . . . . . . . . . . . . . . 60
6.2 Program Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.3 Onshore environments . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.3.1 Examples with Insulation Materials . . . . . . . . . . . . . . . 75
6.3.2 Examples without Insulation Materials . . . . . . . . . . . . . 77
6.4 Oﬀshore Environments . . . . . . . . . . . . . . . . . . . . . . . . . . 79
xii
6.4.1 Examples with Insulation Materials . . . . . . . . . . . . . . . 79
6.4.2 Examples without Insulation Materials . . . . . . . . . . . . . 81
6.5 Adding NonCondensable Gas (N
2
) in an Onshore environment . . . 84
6.5.1 Examples with Insulation Materials . . . . . . . . . . . . . . . 84
6.5.2 Examples without Insulation Materials . . . . . . . . . . . . . 89
6.6 Adding NonCondensable Gas (N
2
) in an Oﬀshore environment . . . 95
6.6.1 Examples with Insulation Materials . . . . . . . . . . . . . . . 95
6.6.2 Examples without Insulation Materials . . . . . . . . . . . . . 99
7 Summary, Conclusions and Future Work 102
7.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
7.2 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
7.3 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
A Derivation of the Equations 110
A.1 Total Energy Equation . . . . . . . . . . . . . . . . . . . . . . . . . . 110
A.2 Mechanical energy balance or the Extended Bernoulli Equation . . . 113
A.3 Evaluation of Heat Loss to the Surrounding . . . . . . . . . . . . . . 115
A.4 Determination of the U
to
and T
ci
. . . . . . . . . . . . . . . . . . . . . 116
A.5 Determination of the Convection Heat Transfer Coeﬃcient . . . . . . 120
A.6 Determination of the Radiation Heat Transfer Coeﬃcient . . . . . . . 122
A.7 Computational Procedure for U
to
. . . . . . . . . . . . . . . . . . . . 122
A.8 Determination of f(t) . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
A.9 Evaluation of the Derrivatives . . . . . . . . . . . . . . . . . . . . . . 127
A.10 Calculation of the Annulus Fluid Properties . . . . . . . . . . . . . . 128
xiii
B Codes for Heat Loss Calculations 132
B.1 Heat Losses from Surface Line . . . . . . . . . . . . . . . . . . . . . . 132
B.2 Heat Losses from Sea Part . . . . . . . . . . . . . . . . . . . . . . . . 137
B.3 Heat Losses from Sea Floor to Reservoir . . . . . . . . . . . . . . . . 145
B.4 Table 14 from Prats [41] . . . . . . . . . . . . . . . . . . . . . . . . . 154
B.5 f(tD) calculation also known as Ramey[42] . . . . . . . . . . . . . . . 155
C Results for Diﬀerent Insulation Materials 156
Bibliography 177
xiv
List of Tables
3.1 Thermal Conductivity of the materials . . . . . . . . . . . . . . . . . 16
3.2 Time Function f(t
D
) for the boundary condition model [49]. . . . . . 25
6.1 Input parameters from Prats [41] as used for diﬀerent example calcu
lations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.2 Radiationnatural convection coeﬃcient of heat transfer. . . . . . . . 57
6.3 Field data parameters for ﬁeld data 1 and ﬁeld data 2 [19]. . . . . . . 63
xv
xvi
List of Figures
1.1 Schematic view of the objective of our calculations, (retrieved from [2]). 4
2.1 Emeraude ﬁeld location and ﬁve spot[9]. . . . . . . . . . . . . . . . . 7
2.2 Schematic view of conduction (after [8]). . . . . . . . . . . . . . . . . 11
2.3 Heat transfer from a hot surface to air by convection (retrieved from
[1]). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 Representation of heat transfer by radiation(after [8]). . . . . . . . . . 12
3.1 Schematic representation of resistance to heat transfer with or without
temperature proﬁle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Schematic representation of resistance to heat transfer with tempera
ture proﬁle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3 Sea water properties change with temperature and salinity[36]. . . . . 21
3.4 Schematic representation of the wellbore. . . . . . . . . . . . . . . . . 21
3.5 Schematic representation of resistance to heat transfer sea ﬂoor to
reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.6 Pressureenthalpy diagram (retrieved from [24]). . . . . . . . . . . . . 26
3.7 Gasliquid ﬂowpatterns for vertical pipes (retrieved from [12]). . . . . 27
3.8 Vertical downward twophase ﬂow [33]. . . . . . . . . . . . . . . . . 29
3.9 Liquid Holdup and Slippage eﬀect representation (retrieved from[4]). 30
3.10 Flow Map for the Beggs and Brill Correlation (retrieved from[4]). . . 34
xvii
3.11 Segregated Flow Regime (retrieved from[4]). . . . . . . . . . . . . . . 35
3.12 Intermittent Flow Regime (retrieved from[4]). . . . . . . . . . . . . . 36
3.13 Distributed Flow Regime (retrieved from[4]). . . . . . . . . . . . . . . 36
3.14 Flow Pattern map for Aziz et al. (retrieved from[14]). . . . . . . . . . 40
5.1 User interface developed GUI for onshore calculations. . . . . . . . . 50
5.2 User interface developed GUI for oﬀshore calculations. . . . . . . . . 51
5.3 User interface developed GUI for both onshore and oﬀshore results. . 51
5.4 User interface developed GUI postprocessing for both onshore and
oﬀshore results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.1 Surface lines heat loss calculation with six diﬀerent insulation materials. 58
6.2 Surface Heat Loss calculation without insulation. . . . . . . . . . . . 58
6.3 Heat loss from sea level to sea ﬂoor with six diﬀerent insulations. . . 59
6.4 Heat loss from sea level to sea ﬂoor without insulation. . . . . . . . . 59
6.5 Heat loss calculation using diﬀerent insulation materials based on Ex
ample 10.2 from Prats[41]. . . . . . . . . . . . . . . . . . . . . . . . . 62
6.6 Heat loss calculation without using insulation materials based on Ex
ample 10.2 from Prats [41]. . . . . . . . . . . . . . . . . . . . . . . . . 62
6.7 Comparison of steam temperature with ﬁeld data 1 and twophase
correlations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.8 Comparison of steam pressure with ﬁeld data 1 and twophase corre
lations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.9 Calculated steam quality with diﬀerent twophase correlations based
on ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.10 Calculated heat loss calculation with insulated tubing based on ﬁeld
data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
xviii
6.11 Comparison of steam temperature with ﬁeld data 2 and twophase
correlations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.12 Comparison of steam pressure with ﬁeld data 2 and twophase corre
lations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.13 Calculated steam quality with diﬀerent twophase correlations based
on ﬁeld data 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.14 Calculated heat loss calculation with insulated tubing based on ﬁeld
data 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.15 Comparison of our model with Fontanilla’s model steam temperature
with ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.16 Comparison of our model with Fontanilla’s model steam pressure with
ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.17 Comparison of our model with Fontanilla’s model steam quality with
ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.18 Comparison of our model with Fontanilla’s model steam temperature
with ﬁeld data 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.19 Comparison of our model with Fontanilla’s model steam pressure with
ﬁeld data 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.20 Comparison of our model with Fontanilla’s model steam quality with
ﬁeld data 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.21 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
using black aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.22 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using
black aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
xix
6.23 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using
black aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.24 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using black
aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.25 Steam temperature distribution , 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . 77
6.26 Steam pressure distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 77
6.27 Steam quality distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 78
6.28 Heat loss distribution, 1 year, T
m
= 122
o
F and injection rate 4850
lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . . . . 78
6.29 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for
black aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.30 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black aerogel. 79
6.31 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black
aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.32 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black aerogel. 80
6.33 Steam temperature distribution, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . 81
xx
6.34 Steam pressure distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 81
6.35 Steam quality distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 82
6.36 Heat loss distribution, 1 year, T
m
= 122
o
F and injection rate 4850
lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . . . . 82
6.37 Pressure drop distribution and formation pressure (green dots), 1 year,
T
m
= 122
o
F and injection rate 4850 lbm/hr with using black aerogel
for onshore. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.38 With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . 84
6.39 With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . . . . 85
6.40 With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr with using black aerogel. . . . . . . . 86
6.41 With changing injection temperature, 1 year, T
m
= 122
o
F and injec
tion rate 4850 lbm/hr with using black aerogel. . . . . . . . . . . . . 87
6.42 With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . . . . 88
6.43 With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . 89
6.44 With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 90
6.45 With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr without insulation. . . . . . . . . . . . 91
6.46 With changing injection temperature, 1 year, T
m
= 122
o
F and injec
tion rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . 93
xxi
6.47 With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 94
6.48 With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . 96
6.49 With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . . . . 96
6.50 With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr with using black aerogel. . . . . . . . 97
6.51 With changing injection temperature, 1 year, T
m
= 122
o
F and injec
tion rate 4850 lbm/hr with using black aerogel. . . . . . . . . . . . . 97
6.52 With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . . . . 98
6.53 With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . 99
6.54 With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 99
6.55 With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr without insulation. . . . . . . . . . . . 100
6.56 With changing injection temperature, 1 year, T
m
= 122
o
F and injec
tion rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . 100
6.57 With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . 101
A.1 Schematic view of tubing element in our calculation. . . . . . . . . . . 111
A.2 Time conduction function (retrieved from[19]). . . . . . . . . . . . . . 126
A.3 Viscosity of the annular ﬂuid with respect to Temperature. . . . . . . 129
A.4 Thermal conductivity of the annular ﬂuid with respect to Temperature. 130
xxii
B.1 Schematic representation of resistance to heat transfer with/without
temperature proﬁle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
B.2 Schematic representation of resistance to heat transfer with tempera
ture proﬁle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
B.3 Schematic representation of resistance to heat transfer sea ﬂoor to
reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
C.1 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft) , 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
white aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
C.2 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white
aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
C.3 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white
aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
C.4 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel. 158
C.5 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
ﬁber glass. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
C.6 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.159
C.7 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.159
C.8 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass. . 160
xxiii
C.9 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
carbon ﬁber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
C.10 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon
ﬁber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
C.11 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon
ﬁber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
C.12 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber. 162
C.13 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
thermolastic insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . 163
C.14 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermo
lastic insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
C.15 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermo
lastic insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
C.16 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic
insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
C.17 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
calcium silicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
xxiv
C.18 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium
silicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
C.19 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium
silicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
C.20 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.166
C.21 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
white aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
C.22 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white
aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
C.23 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white
aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
C.24 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel. 168
C.25 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
ﬁber glass. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
C.26 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.169
C.27 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.170
xxv
C.28 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass. . 170
C.29 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
carbon ﬁber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
C.30 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon
ﬁber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
C.31 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon
ﬁber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
C.32 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber. 172
C.33 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
thermolastic insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . 173
C.34 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermo
lastic insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
C.35 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermo
lastic insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
C.36 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic
insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
xxvi
C.37 Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with
calcium silicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
C.38 Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium
silicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
C.39 Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium
silicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
C.40 Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.176
xxvii
xxviii
Chapter 1
Introduction
Increasing oil prices have helped raise investment in EOR applications during the last
two decades. Steam injection is the thermal method that add heat to the reservoir to
expand the oil inplace, reduce oil viscosity, provide drive energy and thereby improve
the displacement eﬃciency of injected ﬂuid [44] . Crudeoil viscosity is inversely pro
portional to temperature. When the temperature increases, viscosity decreases. Less
viscous oil results in greater mobility. Prats[41] indicated that the eﬀect of steam
injection on recovery is signiﬁcantly greater as compared to hotwater injection. It
is because steam carries more enthalpy per unit mass. The displacement of ﬂuids by
steam is self stabilizing movement of the crude oil to the production well. Lake[30]
emphasized that thermal methods especially steam injection and steam soak, are eas
ily the most successful enhanced oil recovery processes. Steam injection is applied to
viscous oil reservoirs in order to reduce oil viscosity and increase production. Many
applications of steam injection have been done with success in the onshore environ
ment, but only one example exists in the literature for oﬀshore ﬁelds. It was the
”Emeraude Vapeur” pilot test that had great technical success [9].
Many investigators have studied heat transfer to the surrounding formation while
1
2 CHAPTER 1. INTRODUCTION
hot ﬂuid injection travels downward along the wellbore. Ramey [42] was the pio
neer and his model is used by most researchers as a starting point. Ramey assumed
that ﬂow is steady state, noncompressible, and single phase. He solved energy and
momentum balance equations analytically to get approximate pressure and temper
ature distributions. He neglected kinetic energy, frictional loss, and assumed that
the overall heat transfer coeﬃcient does not change with depth. In 1967, Willhite
[49] proposed a well known overall heat transfer coeﬃcient calculation that has been
widely used in the oil industry since. In 1969, Earlougher [17] considered steam and
casing conditions with respect to depth. He used the Hagedorn and Brown model
[22] for twophase ﬂow calculation and assumed that there is no slippage between the
phases for steam injection. He actually assumed that the gas and liquid phase ﬂow
at the same speed. Several years later Pacheco and Farouq Ali [38] presented a com
prehensive mathematical model of steam injection without taking into account slip
and the multiphase ﬂow regime concept. Early in the 1980s, Farouq Ali [18] solved
this issue and proposed a comprehensive wellbore steam ﬂow model. He took into
account the slip concept and ﬂow regime of the ﬂow and concluded that considering
the slip and the ﬂow regime is important for calculating pressure drop and steam
quality during the downward steam injection.
Heat loss and pressure drop calculation for steam injection in oﬀshore environ
ments are not reported in the literature. The objective of this study is to calculate
the heat losses, pressure losses, temperature losses, and steam quality changes along
the wellbore during steam injection in both onshore and oﬀshore environments. For
the onshore cases, pressure drops and quality changes along the wellbore are cal
culated by coupling nonlinear equations and solving them simultaneously. Several
twophase correlations in the literature are used. Oﬀshore cases take into account the
thermophysical properties of sea water to get the correct radiation and convection
heat transfer coeﬃcients. For all calculations, we start with a base case following
3
similar procedures as shown by Fontanilla[20], then extend our work by applying
several more updated twophase correlations. The results obtained using diﬀerent
twophase ﬂow correlations are compared. Sensitivity analysis using diﬀerent insula
tion materials are also conducted to investigate the eﬀect of insulation materials on
downhole steam properties. As another step forward, we conduct calculation of heat
transfer, pressure drops and steam quality change along the wellbore during injection
of steam with an additive of noncondensable gas, such as N
2
. This is the ﬁrst time
that this kind of calculation is described in the open literature. A novel approach is
proposed to solve this problem.
Figure 1.1 shows the schematic view of the our calculation for both onshore and
oﬀshore environments. In order to calculate heat losses from the oﬀshore environ
ments you must consider surface lines, sea level to sea ﬂoor and sea ﬂoor to the
reservoir. Surface lines does not contribute heat losses with comparing sea level and
sea ﬂoor to reservoir, however, insulating small amount of length gives much eﬃciency
on steam quality.
4 CHAPTER 1. INTRODUCTION
Figure 1.1: Schematic view of the objective of our calculations, (retrieved from [2]).
1.1 Thesis Outline
In Chapter 2, we review two cases of heat ﬂow from wells for oﬀshore ﬁelds. One
case had great success in terms of producing more oil resulting from steam injec
tion, the other case used vacuum insulating tubing (VIT) to reduce heat losses and
failed. Then, we review a summary of three basic heat transfer mechanisms and their
combinations that are the backbone of our study. In addition, the wellbore heat
transmission concept is reviewed starting from Ramey’s [42] classical paper.
In Chapter 3, we discuss model formulation that we used in our calculations.
1.1. THESIS OUTLINE 5
Starting with heat transfer calculation in three parts: surface lines, sea, and sea
ﬂoor to reservoir based on both Prats’s [41] and Willhite’s [49] algorithm. Besides
the analogy for oﬀshore, we also provide a robust solution that takes into account
thermo physical properties of seawater in order to calculate important parameters [36]
for heatloss calculations. Continue with steam properties calculation, we present
and discuss most of the twophase correlations that can be applied for downward
steam injection operations. Several of correlations are explained in detail. For other
correlations, references are provided.
In Chapter 4, the problem of steam with noncondensable gas (N
2
) is consid
ered. The N
2
gives additional pressure into our steam and helps to reduce heat loss.
Once N
2
is injected into the system, it causes a decrease in steam pressure and so
steam temperature. In order to achieve this purpose, we calculated steam and non
condensible gas properties for insulated and noninsulated tubing in both onshore
and oﬀshore environments.
Chapter 5, is the one of the exciting parts of this thesis because we developed
a Graphical User Interface (GUI) for our calculations and gave the basics of the
program input and output that is used in our calculations. This program allows a
user to choose several correlations to be applied such as insulated or uninsulated
tubing. With this program a user can study the role of several parameters and see
the eﬀects of those parameter on the system.
Chapter 6 discusses the results obtained in preceding chapters and compares them
in terms of using diﬀerent insulation materials and steam properties. We validated
our program with ﬁeld data from the literature [11] and obtained good agreement
with ﬁeld data and also with Fontanilla’s approach [19]. We augment Fontanilla’s
approach as well.
A summary of our ﬁndings and suggestions for future work are presented in Chap
ter 7.
Chapter 2
Literature Review
We review two examples from the literature for oﬀshore cases of wellbore heat losses.
The ﬁrst is Emeraude Vapeur[9] and subsequently the Marlin failure and redesign
[16, 43, 48] are discussed. This chapter continues with heat transfer mechanism and
with heat transfer discussions.
2.1 Emeraude Vapeur : A Steam Pilot in an Oﬀ
shore Environment
The Emeraude ﬁeld is located oﬀshore Zaire (Congo), Figure 2.1 a, on the West
African coast. Water depth is 65 m. The depth of reservoirs is shallow (200500 m)
and they consist of silt layers alternating with thin fractured limestone beds. These
very heterogeneous reservoirs are signiﬁcantly depleted and oil is viscous (0.1 Pa s
(100 cp)) at reservoir conditions. A steam drive pilot test was decided in order to
estimate a recovery rate and an oilsteam ratio on two independent reservoirs in 1980.
The adverse environmental conditions required original solutions: tilted conductor
pipes, a tilted rig, and adapted pumping units on one platform because of reservoir
6
2.1. EMERAUDE VAPEUR : ASTEAMPILOT INANOFFSHORE ENVIRONMENT7
shallowness and steam production equipment on a second platform because of the
distance to the shore. The Emeraude ﬁeld is estimated to contain 1 billion barrels of
viscous original oil in place (OOIP). After 14 years of production (19721986), only 170
million barrels had been recovered, about 3% of OOIP, and the reservoir was severely
pressure depleted. To produce the remaining reserves by primary recovery in 1520
years, several additional platforms would be needed, and the ﬁnal recovery would still
be only 510% OOIP [9]. Various EOR methods were considered to meet the challenge
(a) Emeraude ﬁeld location. (b) Emeraude ﬁve spot.
Figure 2.1: Emeraude ﬁeld location and ﬁve spot[9].
of this large amount of oil associated with this poor recovery rate. Water injection
was implemented in 1972 in a ﬁvespot pattern, assuming that imbibition would be
active. The results were disappointing, with water breaking through to producers
almost immediately. Insitu combustion tests under laboratory conditions showed
that most of the oil would be burned in the fracture network. Promising results,
however, were obtained from steaminjection tests under laboratory conditions. The
8 CHAPTER 2. LITERATURE REVIEW
experimental work of Willman et al.[15] shows that both steam and hotwater drives
may improve oil mobility by reducing viscosity and also may reduce residual oil at
high temperatures.
A steamﬂood in Figure 2.1 b shows the ﬁve spot steam injection implemented to.
• prove the technological feasibility of such a project in adverse conditions (drilling
with a tilted rig, drilling through and cementing depleted and fractured zones,
pumping tilted wells, and producing steam from seawater) and
• evaluate reservoir responses to steam injection, steam injectivity, steam break
through time, recovery rate, and OSR (oil steam ratio).
The Emeraude steam drive pilot provided original solutions to recover a larger amount
of OOIP than could be recovered by primary production despite diﬃcult conditions.
Technological problems were solved during the pilot design. One of the ﬁrst devel
opments was obtained interms of getting suﬃcient well spacing by using a tilt rig for
drilling. A second success of development was that depleted and fractured zones were
adequately drilled and completed. A third was that pumping units were adapted to
tilted wells and electronically regulated. Lastly, facilities for steam production from
seawater were located on a platform because of the long distance to shore.
In addition to these technological solutions, some reservoir engineering conclusions
were obtained. For instance, steam can be injected at a suﬃcient rate in reservoirs
R1 and R2. A ﬁnal conclusion was that steam improves the oil production rate in
heterogeneous reservoirs. Signiﬁcant response by Well EMV07 in reservoir R2 (oil
rate increased fourfold) and other wells, located on adjacent platforms were noted.
2.2. MARLIN FAILURE ANALYSIS AND REDESIGN 9
2.2 Marlin Failure Analysis and Redesign
In this section, three papers are reviewed related to this subject and including basic
information about Vacuum Insulated Tubing (VIT). Bradford et al. [16] describe
the Marlin failure and give several possible failure modes. They try to reduce them
by making analysis and physical evidence. Ellis et al. [43] apply the failure analysis
from the ﬁrst paper to the remaining Marlin wells and focuses on the VIT redesign
process. Gosch et al. [48] addresses focusing on the value of combined VIT and
ﬁber/software monitoring system as a means of both controlling and observing well
thermal behavior.
VIT provides a solution for heat loss in steam injection in Arctic and oﬀshore
environments. VIT is a tubular apparatus conveying steam or other hot ﬂuids (>
400
o
F) to the formation through an inner tubular that is surrounded by an outer
tubular [6]. The annular space between the inner and the outer tubular is under
vacuum. The annulus is usually ﬁlled with a better material for absorbing hydrogen
formed by corrosion of the outer tubular and gases such as nitrogen, carbon monoxide,
and hydrogen released from the inner tubular at elevated temperatures [6].
The Marlin ﬁeld is located in the Gulf of Mexico, Viosca Knoll Blocks 871/915,
and was originally intended to be produced from a tensionlog platform by means of
ﬁve predrilled drytree penetrations. Right after production began, a minor tubing
leak occurred, and after that casing pressure jumped to shut in tubing pressure. It
was the ﬁrst alert of a major tubing failure. Analysis of the failure came up with
several critical issues including
• excessive helical buckling of the production tubing,
• hydrate formation and dissolution,
• trapped annulus pressure leading to casing collapse, and
10 CHAPTER 2. LITERATURE REVIEW
• improper tubulars and wellhead movement
After the ﬁshing operation and ultrasonic caliper, they concluded that deformation
of Well A2’s tubing was the result of collapse of production tieback. They still do
not know whether it was because of the production tieback collapsed alone or as a
result of collapse of the intermediate casing is unknown [16].
Although Marlin wells predrilled up to the completion stage limited mitigation
options, well design concepts were developed and screened using agreedupon risk
acceptance criteria for health, safety, and environment (HSE). They chose VIT based
on the economic analysis and risk proﬁle associated with each option. Using VIT,
they had signiﬁcant thermal isolation due to the low conductivity or low convecting
annular ﬂuids [43]. They did several experimental studies on VIT and found out
three important facts. The overall thermal properties of the unaltered VIT were
not adequate for Marlin requirements due to the fact that heat loss at coupling
dominates the performance of a VIT joint. Second, a combination of thermal coatings
and insulated inserts provides adequate additional insulation at the couplings. Last,
several materials were tested in the annulus to reduce heat loss and N
2
is the most
eﬀective barrier to heat loss.
In the third paper of the Marlin failure redesign [48], they have tested VIT per
formance with developed software to see the production annulus temperature proﬁle
in real time. This realtime monitoring gave them better control of the well such
as when a low safety factor is calculated, a well is shut in automatically. VIT de
sign itself has lots of design considerations and challenges. They tried to understand
whether the VIT was a good choice or not based on the both experimental and nu
merical results [48]. They sum up their work in several with giving results. Design of
VIT introduces a number of considerations not present in a design using conventional
tubing. Another important conclusion was natural convection can signiﬁcantly aﬀect
the ability of VIT to isolate tubing temperatures from the annulus. Additionally, it
2.3. HEAT TRANSMISSIONMECHANISMS ANDDISCUSSIONFROMAUTHORS11
was remarked that regional heating can add a temperature increase to outer annuli
that is not anticipated in a single well analysis. Both the thermal performance of the
VIT and its mechanical integrity require special consideration. These results lead the
authors to suggest that each well has to studied and treated individually.
2.3 Heat Transmission Mechanisms and Discus
sion from Authors
2.3.1 Heat Transmission Mechanisms
In this section we review the heat transfer mechanisms during steam injection oper
ations. Including conduction, convection, and radiation and a combination of two or
more in our calculations.
2.3.1.1 Heat Transfer by Conduction
Figure 2.2: Schematic view of conduction
(after [8]).
Heat conduction also called diﬀusion is
the transfer of energy from the more en
ergetic particles of a substance to the ad
jacent, less energetic ones as a result of
interaction between particles.
˙
Q
cond
= kA
T
1
−T
2
∆x
= −kA
∆T
∆x
(2.1)
where k is the thermal conductivity of
the material (Fig. 2.2). In the limit
ing case ∆x → 0 the equation above
reduces to the diﬀerential form that is
12 CHAPTER 2. LITERATURE REVIEW
called Fourier
s law of heat conduction after J. Fourier and becomes [8, 35].
˙
Q
cond
= −kA
dT
dx
(2.2)
2.3.1.2 Heat Transfer by Convection
Figure 2.3: Heat transfer
from a hot surface to air
by convection (retrieved
from [1]).
One mode of convection heat transfer is between a solid
surface and the adjacent liquid or gas that is in motion.
Convection is called forced convection if the ﬂuid is forced
to ﬂow over the surface by external means such as fan,
pump, or the wind. In contrast, convection is called natu
ral (or free) convection if the ﬂuid motion caused by buoy
ancy forces that are induced by density diﬀerences due to
the variation of temperature in the ﬂuid (Fig. 2.3). The
rate of convection heat transfer is expressed by Newton
s
law of cooling as [8, 35]
˙
Q
conv
= −hA
s
(T
s
−T
∞
) (2.3)
2.3.1.3 Heat Transfer by Radiation
Figure 2.4: Representation
of heat transfer by radia
tion(after [8]).
Radiation is the energy emitted by matter in the form
of electromagnetic waves as a result of changes in the
electronic conﬁgurations of the atoms or molecules.
The radiation that can be emitted from a surface at
an absolute temperature T
s
(in K or R) is given by
the Stefan −Boltzman law as [8, 35]
˙
Q
emit,max
= σA
s
T
4
s
(2.4)
2.3. HEAT TRANSMISSIONMECHANISMS ANDDISCUSSIONFROMAUTHORS13
2.3.2 Heat Transmission Discussion from Authors
In the literature many investigators have worked on the thermodynamic properties of
the hot ﬂuid movement through wells in onshore ﬁelds for both production and injec
tion. One of the wellknown papers for wellbore heat transmission is by Ramey[42].
Most of the publications follow upon his approach. He provided an approximate ana
lytical solution for wellbore heat transmission. In his paper, Ramey [42] made several
assumptions. He assumed that ﬂuid is noncompressible and ﬂow is single phase with
constant thermal and physical properties along the wellbore. He considered that heat
ﬂows radially away the wellbore and the overall heat treansfer coeﬃcient is indepen
dent of depth. He did not take into account frictional pressure loss and kinetic energy
eﬀect in his calculation.
Squier et al. [47] solved diﬀerential equations describing ﬂuid temperature along
the wellbore, using a complete analytical method. They used hot water as injection
ﬂuid.
Satter [45] presented a method that improved Ramey’s [42] model by making
the overall heat transfer coeﬃcient dependent on depthstep method for calculating
heat loss and steam quality for saturated steam as a function of depth. Since Satter
assumed that there is no change in pressure with depth, he assumed, in eﬀect that
the temperature of the injected saturated steam remains constant, and that only the
quality varies with depth.
One year after Satter’s [45] paper, Holst and Flock [25] added the friction loss and
kinetic energy eﬀects on Ramey’s [42] and Satter’s [45] models, in order to calculate
the heat loss and quality distribution versus depth for saturated steam injection
operations. They neglected, however, the static pressure change.
In 1966, Leutwyler [31] gave a comprehensive treatment of casing temperature
behavior. Hans and Huitt [26] also developed a graphical solution for wet steam
14 CHAPTER 2. LITERATURE REVIEW
injection operations. In their model, they calculate wellbore heat loss, steam conden
sation rate, and casing temperature.
One year later, Willhite [49] proposed his well known method for estimation of
overall heat transfer coeﬃcient that is applied in our calculation as well.
Two of the pioneers in the prediction of heat loss and pressure drop in the wellbore
were Pacheco and Farouq Ali [38]. They formulated a mathematical model that
consisted of two coupled nonlinear diﬀerential equations that were solved iteratively
in terms of pressure and quality of steam. They assumed single phase ﬂow, which is
not valid, and later on Farouq Ali [18] solved this problem by taking into account slip
between the ﬂuids and the ﬂow regime. He used several correlations and stated that
importance of applying twophase ﬂow concept and ﬂow regime.
Wu and Pruess [50] presented a new analytical for wellbore heat transmission
without Ramey’s assumptions. Their approach was assuming nonhomogeneous for
mations as layered formation with diﬀerent physical properties.
All the development both technological and understanding of the problem have
been done so far helped to understand heat transfer mechanism and solve the problem
with taking into account considerations from the authors.
Chapter 3
Model Formulation
This chapter presents the methodology for heatloss calculations and pressure drop
estimation in injection tubing.
3.1 Heat Loss Calculations
We have adapted Fontanilla’s [19] assumptions for the solution of oﬀshore and onshore
injectors. These assumptions follow:
1. Heat transfer in and around the wellbore takes place under pseudo steady state
conditions. Conduction is pseudo steady state when the change in temperature
with respect to time at any given point in the wellbore remains constant. This
would mean that the heat rate through the wellbore components is the same at
any given time. This can be seen in the derivation of the equation to evaluate
the overall heat transfer coeﬃcient U
to
. The heat transfer into the earth occurs
under unsteadystate condition.
2. The heat diﬀusivity and the conductivity of the formation is constant.
3. The tubing hangs symmetrically inside the casing.
15
16 CHAPTER 3. MODEL FORMULATION
As steam travels downward in the wellbore, it loses energy to the surrounding
formation. This heat loss may result in condensation with consequent reduction in
steam quality and enthalpy.
Heat losses through pipes, whether surface lines or wells, usually are estimated
at steadystate conditions in oilﬁeld operations. The procedures for estimating heat
losses may appear laborious. The calculation procedures of both Prats [41] and
Willhite [49] are discussed here. For oﬀshore heat losses calculations an analogy
between electrical circuits and heat resistance is made.
One of the classic papers about overall heat transfer coeﬃcient was published
by Willhite [49]. He presented his widely used method that is incorporated in most
simulators for hot ﬂuid injection and hot water. Here in our study, we will give the
equations for both surface lines and wellbores based on the methods discussed. For
comprehensive calculation procedure we refer to read Appendix A. In addition, we
provide the equations for insulated and noninsulated cases and also for oﬀshore heat
loss calculations. In Table 3.1, we have provided thermal conductivity of the diﬀerent
insulation material based on the values from the thesis of Marques[34].
Table 3.1: Thermal Conductivity of the materials
Insulation Materials W/(m*K) (BTU/(ft −hr −
o
F))
Black Aerogel 0.012 0.0069
White Aerogel 0.014 0.0081
Fiberglass 0.028 0.0162
Carbon Fiber 0.036 0.0208
Thermolastic Insulation 0.041 0.0237
Calcium Silicate 0.069 0.0400
3.1. HEAT LOSS CALCULATIONS 17
3.1.1 Heat Loss from Surface Lines
3.1.1.1 With/without Insulation
(a) Surface pipe with insulation. (b) Surface pipe without insulation.
Figure 3.1: Schematic representation of resistance to heat transfer with or without
temperature proﬁle.
In Figure 3.1 a. we have a representative resistance to heat transfer and a temperature
proﬁle. Prats [41] stated that even though heat losses from surface lines in hot ﬂuid
injection operations may be a small fraction of the total heat injected, it is generally
worthwhile to use insulation to reduce heat losses to save both money and fuel. That
savings can be signiﬁcant will be demonstrated by means of steam injection examples.
The basic equation used to calculate heat losses per unit length of pipe,
˙
Q
ls
, is
˙
Q
ls
=
T
b
−T
A
R
h
(3.1)
where R
h
is represented as R
h
=
1
2πrU
, U is the overall coeﬃcient of heat transfer, and
18 CHAPTER 3. MODEL FORMULATION
r is an arbitrary radius that usually coincides with the radius of one of the surfaces for
which the heat loss is being determined. Here, R
h
is the speciﬁc thermal resistance
(thermal resistance per unit length of pipe) and is given in units of (BTU/(ft −hr −
o
F))
−1
, T
b
is the bulk temperature of the ﬂuid in the pipe in degrees Fahrenheit, T
A
is the ambient temperature of the atmosphere in degrees Fahrenheit, and
˙
Q
ls
is the
rate of heat loss per unit length of pipe in BTU/(ft −hr). Rates of heat loss during
transient periods can be several times greater than at steady state. Transient eﬀects
generally are neglected in calculations of heat losses from thermal lines, because the
transient phase is often of short duration (of the order of less than a day). For a pipe
covered with insulation, the speciﬁc thermal resistance of heat loss is given as
R
h
=
1
2π
_
1
h
f
r
i
+
1
h
pi
r
i
+
1
λ
p
ln
r
o
r
i
+
1
h
po
r
o
+
1
λ
ins
ln
r
ins
r
o
+
1
h
fc
r
ins
_
(3.2)
Here h
f
is the ﬁlm coeﬃcient of heat transfer between the ﬂuid inside the pipe and
the pipe wall, h
pi
is the coeﬃcient of heat transfer across any deposits of scale or
dirt at the inside wall of the pipe, h
po
is the coeﬃcient of heat transfer across the
contact between pipe and insulation, h
fc
is the coeﬃcient of heat transfer due to
forced convection (air currents) at the outer surface of the insulation, r
i
is the inner
radius of the pipe,r
o
is the outer radius of the pipe and essentially the inner radius
of the insulation, r
ins
is the external radius of the insulation, and λ
p
and λ
ins
are the
thermal conductivities of the pipe and insulation. Coeﬃcients of heat transfer are
expressed in (BTU/(sqft − hr −
o
F)) , radii in feet, and thermal conductivities in
(BTU/(ft − hr −
o
F)). Because the temperature on the surface of most insulated
lines is low, radiation is usually insigniﬁcant and is not included in Eq.3.2 above. The
physical signiﬁcance of each of the six terms in the right side of Eq.3.2 is illustrated
in Figure 3.1. Each of the six terms is proportional to a thermal resistance in the
system aﬀecting heat losses. Adjacent to the inner surface of the pipe is a lowvelocity
3.1. HEAT LOSS CALCULATIONS 19
ﬂuid ﬁlm (1). Because of its low velocity, this ﬁlm has heat transfer characteristics
diﬀerent from those of the ﬂowing bulk ﬂuid and accounts for the introduction of
the ﬁlm coeﬃcient of heat transfer h
f
. Note that the resistance to heat ﬂow across
this ﬁlm decreases as the value of the coeﬃcient of heat transfer increases. Scale or
dirt deposits at the inside (2) and outside (4) pipe walls lead to coeﬃcients of heat
transfer h
pi
and h
po
, respectively. Heat transfer through the pipe wall (3) and the
insulation (5) is by conduction. A low velocity ﬂuid ﬁlm at the exterior surface of
the insulation (6), which aﬀects heat losses to the atmosphere by forced convection,
leads to the coeﬃcient of heat transfer h
fc
. It should be pointed out that adding
more insulation does not necessarily reduce the rate of heat losses further.
3.1.2 Heat Loss from Sea Level to Sea Floor
3.1.2.1 With/without Insulation
(a) Sea Part with insulation. (b) Sea Part without insulation.
Figure 3.2: Schematic representation of resistance to heat transfer with temperature
proﬁle.
20 CHAPTER 3. MODEL FORMULATION
For the sea level heat loss calculation, we have made several assumptions taking
Figure 3.4 as reference:
1. heat transfer is steady because the speciﬁed thermal conditions at the bound
aries do not change with time,
2. heat transfer is one dimensional due to thermal symmetry about the midpoint,
3. thermal conductivity is constant,
4. sea temperature does not change along the wellbore
R
total
= R
conv,1
+ R
cycl,1
+R
cycl,2
+R
cycl,3
+R
conv,2
(3.3)
R
total
=
1
h
1
A
1
+
1
2πLk
1
ln
r
2
r
1
+
1
2πLk
2
ln
r
3
r
2
+
1
2πLk
3
ln
r
4
r
3
+
1
h
2
A
4
(3.4)
In order to calculate heat losses from oﬀshore wells we have to ﬁnd T
ins
and U
to
and
also sea water parameters such as thermal diﬀusivity, thermal conductivity, density,
speciﬁc heat at the given salinity and temperature of the sea.
3.1. HEAT LOSS CALCULATIONS 21
Figure 3.3: Sea water properties change with temperature and salinity[36].
Figure 3.4: Schematic representation of the wellbore.
22 CHAPTER 3. MODEL FORMULATION
3.1.3 Heat Loss from Sea Floor to Reservoir
3.1.3.1 With/without Insulation
(a) Sea Floor to Reservoir with insulation. (b) Sea Floor to Reservoir without insulation.
Figure 3.5: Schematic representation of resistance to heat transfer sea ﬂoor to reser
voir.
Heat losses from wells never reach a steady state. They attain, as pointed out by
Ramey [42] and Willhite [49], a quasisteady state in which the rate of heat loss is a
monotonically decreasing function of time. This function of time, is discussed later in
more detail. It is a measure of how fast the earth conducts heat away from the well.
Heat losses from the well to the earth are characterized by Eq. 3.1, where in this
case the ambient temperature is the geothermal temperature and, thus, a function
of depth. In this case, of course, the speciﬁc thermal resistance is time dependent,
reﬂecting the variable eﬀective thermal resistance of the earth. A representation of
the typical elements oﬀering resistance to heat losses from the wellbore is given in
Fig 3.5. For the insulated tubing held concentrically within the casing shown in this
3.1. HEAT LOSS CALCULATIONS 23
ﬁgure, the heat resistance elements are combined to obtain the overall coeﬃcient of
heat loss:
R
h
=
1
2π
_
1
h
f
r
i
+
1
h
pi
r
i
+
1
λ
p
ln
r
o
r
i
+
1
h
po
r
o
+
1
λ
ins
ln
r
ins
r
o
+
1
h
ς,an
r
ins
+
1
λ
p
ln
r
co
r
ci
+
1
λ
cem
ln
r
w
r
co
+
1
λ
Ea
ln
r
Ea
r
w
+
f(t
D
)
λ
E
_
(3.5)
The ﬁrst ﬁve terms have been discussed in the preceding section for heat loss from
surface lines. The last ﬁve terms represent, in order of appearance, the resistance
to radiation and convection in the annulus, the resistance of casing, the resistance
of the cement, the resistance of an altered zone (resulting from drying due to high
temperatures) in the earth, and the variable resistance of the earth.
Diﬀerent well designs lead to diﬀerent expressions for determining the overall
thermal resistance R
h
. In Eq, h
ς,an
is the radiation and convection coeﬃcient of
heat transfer for the annulus, r
ci
and r
co
are the inner and outer casing radii, r
w
is the wellbore radius, r
Ea
is the radius of the altered zone in the earth near the
wellbore, λ
cem
is the thermal conductivity of the cement, λ
Ea
and λ
E
are the thermal
conductivities of the altered and unaltered earth, and f(t
D
) is the time function
that reﬂects the thermal resistance of the earth. Coeﬃcients of the heat transfer
are expressed in (BTU/(sqft −hr −
o
F)), radii in feet, and thermal conductivities in
(BTU/(ft−hr−
o
F)). The function f(t
D
) is dimensionless, and the dimensionless time
is discussed later. The function f(t
D
) and the radiationconvection coeﬃcient of heat
transfer in the annulus, h
ς,an
, are the only additional terms requiring discussion. The
function f(t
D
) has been discussed by a number of authors, here only the Ramey[42]
and Willhite[49] representation is going to be discussed. f(t
D
) is represented in terms
of dimensionless time:
t
D
=
α
E
t
r
2
w
(3.6)
24 CHAPTER 3. MODEL FORMULATION
if there is an altered zone,
t
D
=
α
E
t
r
2
Ea
(3.7)
here α
E
is the thermal diﬀusivity of the earth in square feet per hour, and t is the
time from start of heating in hours. For values of t
D
≤ 100 and is then Ramey [42]
gives calculation:
f(t
D
)
1
2
∗ (lnt
D
) + 0.403 (3.8)
for t
D
≤ 100. For values of of t
D
≤ 100, Willhite [49] has published Table 3.2. This
Table 3.2 is used to interpolate the value for ﬁnding f(t
D
).
3
.
1
.
H
E
A
T
L
O
S
S
C
A
L
C
U
L
A
T
I
O
N
S
2
5
Table 3.2: Time Function f(t
D
) for the boundary condition model [49].
t
D
100 50 20 10 5.0 2.0 1.0 0.5 0.2 0.1 0.05 0.02 0.01 0.0
0.1 0.313 0.313 0.314 0.316 0.318 0.323 0.330 0.345 0.373 0.396 0.417 0.433 0.438 0.445
0.2 0.423 0.423 0.424 0.427 0.430 0.439 0.452 0.473 0.511 0.538 0.568 0.572 0.578 0.588
0.5 0.616 0.617 0.619 0.623 0.629 0.644 0.666 0.698 0.745 0.772 0.790 0.802 0.806 0.811
1.0 0.802 0.803 0.806 0.811 0.820 0.842 0.872 0.910 0.958 0.984 1.000 1.010 1.010 1.020
2.0 1.020 1.020 1.030 1.040 1.050 1.080 1.110 1.150 1.200 1.220 1.240 1.240 1.240 1.250
5.0 1.360 1.370 1.370 1.380 1.400 1.440 1.480 1.520 1.560 1.570 1.580 1.590 1.590 1.590
10.0 1.650 1.660 1.660 1.670 1.690 1.730 1.770 1.810 1.840 1.860 1.860 1.870 1.870 1.880
20.0 1.960 1.970 1.970 1.990 2.000 2.050 2.090 2.120 2.150 2.160 2.160 2.170 2.170 2.170
50.0 2.390 2.390 2.400 2.420 2.440 2.480 2.510 2.540 2.560 2.570 2.570 2.570 2.580 2.580
100.0 2.730 2.730 2.740 2.750 2.770 2.810 2.840 2.860 2.880 2.890 2.890 2.890 2.890 2.900
26 CHAPTER 3. MODEL FORMULATION
3.2 Steam Phase behavior calculations
Figure 3.6: Pressureenthalpy diagram (retrieved from [24]).
The steam properties such as density of the saturated steam and density of the sat
urated liquid are calculated directly using IAPWS IF97[24]. Based on the function
from[24], our pressure versus enthalpy diagram looks like Figure 3.6. Steam quality
changes with depth. The appropriate ordinary diﬀerential equations are described in
Appendix A. To solve this diﬀerential equation in each interval of the well, a fourth
order RungeKutta method is used with Matlab ”ode45” function.
3.3. TWO PHASE FLOW CORRELATIONS 27
3.3 Two Phase Flow Correlations
Unlike singlephase ﬂow, twophase ﬂow behavior is more complex than for single
phase ﬂow. The phases tend to separate because of diﬀerences in density. Shear
stresses at the pipe wall are diﬀerent for each phase because of their diﬀerent densities
and viscosities. The main diﬀerence between gas and liquid phase is they do not
travel at the same speed in the pipe. For downward ﬂow, liquid always ﬂows faster
than the gas or vapor phase. We give information about the two phase correlations
that are applied in our calculations for vertical downward ﬂow with insulated and
uninsulated tubing for both an onshore and oﬀshore environments. The twophase
ﬂow correlations we used in our calculations are modiﬁed Beggs and Brill, Aziz, Govier
and Fogarasi, Drift Flux model, and Hasan and Kabir correlations. Besides, we also
addressed ﬂow regimes for vertical ﬂow. We provide ﬂow regimes in two phase vertical
ﬂow in Figure 3.7.
Figure 3.7: Gasliquid ﬂowpatterns for vertical pipes (retrieved from [12]).
28 CHAPTER 3. MODEL FORMULATION
The general pressure gradient equation is
dp
dz
=
g
g
c
ρ
s
sinφ +
fρ
f
v
2
m
2g
c
d
+
ρv
m
dv
m
g
c
dz
(3.9)
where ρ
s
= ρ
l
H
L
+ ρ
g
H
g
and the deﬁnition for ρ
s
and the density term is used in
the acceleration component.
dp
dz
=
_
dp
dz
_
el
+
_
dp
dz
_
f
+
_
dp
dz
_
acc
(3.10)
The pressure drop caused by elevation change depends on the density of the twophase
mixture and is usually calculated using a liquid holdup value. Friction losses require
evaluation of a twophase friction factor. Acceleration is sometimes negligible and is
usually calculated only for high ﬂow velocities. Many correlations have been developed
for predicting twophase ﬂowing pressure gradients that diﬀer in the manner used to
calculate these three components of the total pressure gradient. Some investigators
chose to assume that gas and liquid phases travel at the same velocity (no slippage
between phases) for evaluating the mixture density and evaluate only a friction factor
empirically. Others developed methods for calculating both liquid hold up and friction
factor and some chose divide the ﬂow conditions into patterns and developed separate
correlations for each ﬂow regime.
Figure 3.8 shows ﬂow regime pattern both injection and production of the ﬂuid.
Predicting the ﬂow regimes that occur at a given location in a well is extremely im
portant. The empirical correlation or mechanistic model used to predict ﬂow behavior
varies with ﬂow pattern. Droplet ﬂow, also known as mist ﬂow, happens at the well
head during steamonly injection. In this ﬂow, the gas phase is continuous and the
bulk of the liquid is entrained as droplets in the gas phase. The pipe wall is coated
with a liquid ﬁlm, but the gas phase predominantly controls the pressure gradient
[13, 14, 23, 46].
3.3. TWO PHASE FLOW CORRELATIONS 29
Annular mist ﬂow also as known as annular droplet ﬂow occurs right after the
mist ﬂow and is characterized by the axial continuity of the gas phase in a central
core with the liquid ﬂowing downward, both as a thin ﬁlm along the pipe wall and as
dispersed droplet in the core. [13, 14, 23, 46].
Figure 3.8: Vertical downward
twophase ﬂow [33].
Churn ﬂow is the change of continuous gas
phase to continuous liquid phase. There is clear
distinction between gas bubbles and liquid phase
like gas phase trapped into largebubbles. Neither
phase appears to be continuous. [13, 14, 23, 46].
During slug ﬂow, the pipe is almost com
pletely ﬁlled with the liquid and free gas phase
is present in small bubbles. The bubbles move at
diﬀerent velocities and except for their density,
have little eﬀect on the pressure gradient. The
wall of the pipe always contacts with the liquid
phase [13, 14, 23, 46].
The correlations addressed in this study are
discerned from each other by taking into account
both the slippage eﬀect and ﬂow patterns. Pa
rameters are calculated in twophase ﬂow requires
knowledge of several parameters such as liquid
holdup, superﬁcial velocity of both gas and liq
uid phases, viscosity of both phases, slip velocity
and noslip velocity values. Here we are going to
give equations of them.
Liquid Holdup and Slippage Eﬀect
When two or more phases are present in a pipe, they tend to ﬂow at diﬀerent
30 CHAPTER 3. MODEL FORMULATION
insitu velocities. These insitu velocities depend on the density and viscosity of each
phase. Typically the phase that is less dense ﬂows faster than the other. This causes
a ”slip” eﬀect between the phases. As a consequence, the insitu volume fractions of
each phase (under ﬂowing conditions) diﬀer from the input volume fractions of the
pipe. Liquid holdup is deﬁned as the ratio of the volume of a pipe segment occupied
by liquid to the volume of the pipe segment. That is
H
L
=
volume of liquid in a pipe segment
volume of pipe segment
(3.11)
Liquid holdup is a fraction that varies from zero for all gas ﬂow to one for all liquid
ﬂow. The remainder of the pipe segment is of course occupied by gas, that is referred
to as gas holdup or gas fraction. That is
H
g
= 1 −H
L
(3.12)
Figure 3.9: Liquid Holdup and Slippage eﬀect representation (retrieved from[4]).
No −Slip Liquid Holdup
Noslip holdup, sometimes called input liquid content, is deﬁned as the ratio of the
volume of liquid in a pipe segment divided by the volume of the pipe segment which
would exist if the gas and liquid traveled at the same velocity (noslippage). It can
3.3. TWO PHASE FLOW CORRELATIONS 31
be calculated directly from the known liquid and gas from rates as
C
L
=
q
L
q
L
+q
g
(3.13)
C
g
= 1 −C
L
=
q
g
q
L
+q
g
(3.14)
Velocity
Many twophase ﬂow correlations are based on a variable called superﬁcial velocity.
The superﬁcial velocity of a ﬂuid phase is deﬁned as the velocity which that phase
would exhibit if it ﬂowed through the total cross section of the pipe alone. Superﬁcial
velocity for the gas phase is
v
sg
=
q
g
A
(3.15)
Actual gas velocity is
v
g
=
q
g
AH
g
(3.16)
Superﬁcial velocity for liquid phase is
v
sL
=
q
L
A
(3.17)
The actual liquid velocity is
v
s
=
q
L
AH
L
(3.18)
The twophase mixture velocity is
v
m
= v
sL
+ v
sg
(3.19)
The slip velocity is
v
s
= v
g
−v
L
=
v
sg
H
g
−
v
sL
H
L
(3.20)
32 CHAPTER 3. MODEL FORMULATION
The noslip holdup is
C
L
=
v
sL
v
m
(3.21)
Viscosity
Viscosity of the both saturated water and saturated steam is calculated based on the
correlation of from Liang et al. [32].
For saturated water viscosity
µ
w
= exp[0.484045 −3.1115 ∗ 10
−2
∗ T
0.95
+ 1.3192 ∗ 10
−4
∗ T
1.9
−2.2934 ∗ 10
−7
∗ T
2.85
¸
(3.22)
For saturated steam viscosity
µ
s
= 0.0085 + exp[−7.0661 + 2.1106 ∗ 10
−2
∗ T −7.2085 ∗ 10
−5
∗ T
2
+1.0111 ∗ 10
−7
∗ T
3
¸
(3.23)
3.3.1 Modiﬁed Beggs and Brill Model
For multiphase ﬂow, many of the published correlations are applicable for ”vertical
ﬂow” only, while others apply for ”horizontal ﬂow” only. Not many correlations
apply to the whole spectrum of ﬂow situations that may be encountered in oil and
gas operations, namely uphill, downhill, horizontal, inclined and vertical ﬂow. The
Beggs and Brill [3, 7, 14, 39] correlation, is one of the few published correlations
capable of handling all these ﬂow directions. It was developed using 1” and 11/2”
sections of pipe that could be inclined at any angle from the horizontal.
The Beggs and Brill multiphase correlation deals with both the friction pressure
loss and the hydrostatic pressure diﬀerence. First the appropriate ﬂow regime for
3.3. TWO PHASE FLOW CORRELATIONS 33
the particular combination of gas and liquid rates (Segregated, Intermittent or Dis
tributed) is determined. The liquid holdup, and hence, the insitu density of the gas
liquid mixture is then calculated according to the appropriate ﬂow regime, to obtain
the hydrostatic pressure diﬀerence. A twophase friction factor is calculated based
on the ”input” gasliquid ratio and the Moody friction factor table using Colebrook
equation. From this the friction pressure loss is calculated using ”input” gasliquid
mixture properties.
3.3.1.1 FlowPattern Determination
The Beggs and Brill correlation needs to identify the ﬂow pattern at the given ﬂowing
conditions in order to calculate the liquid holdup and friction pressure drop. For this
purpose, the Beggs and Brill correlation makes use of a horizontal ﬂow pattern map
built based on the Froude number of the mixture (F
rm
) and input liquid content
(noslip liquid holdup, C
L
).
In order to build the ﬂow map, the observed ﬂow patterns were grouped as: seg
regated (stratiﬁed, wavy and annular ﬂow), intermittent (plug and slug ﬂow), dis
tributed (bubble and mist ﬂow), and transition (ﬂow pattern included after a modi
ﬁcation of the original publication that considers the region between the segregated
and intermittent grouped patterns).
The boundaries between these groups of ﬂow patterns appear as curves in a log
log plot in the original publication by Beggs and Brill. This was later revised so
that straight lines could be used instead. We use this modiﬁed ﬂow pattern map in
our calculations. The revised lines that deﬁne the boundaries are deﬁned as follows
(where * stands for the modiﬁcation of the original curve to a straight line in a loglog
plot)
34 CHAPTER 3. MODEL FORMULATION
Figure 3.10: Flow Map for the Beggs and Brill Correlation (retrieved from[4]).
L
∗
1
= 316C
0.302
L
(3.24)
L
∗
2
= 0.0009252C
−2.4684
L
(3.25)
L
∗
3
= 0.1C
−1.4516
L
(3.26)
L
∗
4
= 0.5C
−6.738
L
(3.27)
The identiﬁed ﬂow pattern is the one that would exist if the pipe were horizontal.
Unless the pipe is actually in the horizontal position, the Beggs and Brill correlation
is not able to recognize the actual ﬂow pattern under the given conditions. Therefore,
to calculate the liquid holdup, we ﬁrst determine the liquid holdup for the horizontal
3.3. TWO PHASE FLOW CORRELATIONS 35
ﬂow, and this value is then corrected for the angle of interest.
The Froude number is a dimensionless number that relates the inertia with respect
to the gravitational forces. For a mixture, it is obtained as
Fr
m
=
V
2
m
gD
(3.28)
Once the input liquid content (C
L
) and Froude number of the mixture (F
rm
) are de
termined, the corresponding ﬂow pattern is identiﬁed when the following inequalities
are satisﬁed.
SegregatedFlow occurs when C
L
< 0.01 and Fr
m
< L
∗
1
or C
L
≥ 0.01 and Fr
m
< L
∗
2
.
Figure 3.11: Segregated Flow Regime (retrieved from[4]).
IntermittentFlow occurs when 0.01 ≤ C
L
< 0.4 and L
3
< Fr
m
≤ L
∗
1
or C
L
≥ 0.4
and L
∗
3
< Fr
m
≤ L
∗
4
.
36 CHAPTER 3. MODEL FORMULATION
Figure 3.12: Intermittent Flow Regime (retrieved from[4]).
DistributedFlow occurs when C
L
< 0.4 and Fr
m
≥ L
∗
1
or C
L
≥ 0.4 and Fr
m
> L
∗
4
.
Figure 3.13: Distributed Flow Regime (retrieved from[4]).
TransitionFlow occurs when C
L
≥ 0.01 and L
∗
2
< Fr
m
< L
∗
3
.
3.3.1.2 Hydrostatic Pressure Diﬀerence
Once the ﬂow pattern has been determined, the liquid holdup is then calculated.
Beggs and Brill divided the liquid holdup calculation into two parts. First, the liquid
holdup for horizontal ﬂow, E
L
(0), is determined. Afterwards, this horizontal holdup
is corrected for inclined ﬂow to obtain the actual holdup, E
L
(θ). The horizontal
3.3. TWO PHASE FLOW CORRELATIONS 37
holdup must be E
L
(0) ≥ C
L
. Therefore, in the event that E
L
(0) < C
L
, the horizontal
holdup is set to E
L
(0) = C
L
. The expression used to calculate the horizontal holdup
changes per ﬂow pattern group as follows in MatLab:
1: if strcmp(Flowpattern,
Segregated
) then
2: E
L
(0) ←
0.98C
0.4846
L
Fr
0.0868
m
3: else
4: if strcmp(Flowpattern,
Intermittent
) then
5: E
L
(0) ←
0.845C
0.5351
L
Fr
0.0173
m
6: end if
7: if strcmp(Flowpattern,
Distributed
) then
8: E
L
(0) ←
1.065C
0.5824
L
Fr
0.0609
m
9: end if
10: if strcmp(Flowpattern,
Transition
) then
11: A ←
L
∗
3
−Frm
L
∗
3
−L
∗
2
12: B ←1 −A
13: E
L
(0)
Segregated
←
0.98C
0.4846
L
Fr
0.0868
m
14: E
L
(0)
intermittent
←
0.845C
0.5351
L
Fr
0.0173
m
15: E
L
(0)
Transition
←AE
L
(0)
Segregated
+BE
L
(0)
Intermittent
16: end if
17: end if
Once the horizontal insitu liquid volume fraction is determined, the actual liquid
volume fraction is obtained by correcting E
L
(0) by an inclination factor B(θ):
E
L
(θ) = B(θ)E
L
(0) (3.29)
B(θ) = 1 +β
_
sin(1.8θ) −
1
3
sin
3
(1.8θ)
_
(3.30)
β for all type of ﬂow pattern is
β = (1 −C
L
)ln
_
4.70N
0.1244
vl
C
0.3692
L
Fr
0.0978
m
_
(3.31)
38 CHAPTER 3. MODEL FORMULATION
where N
v
l = 1.938V
sl
_
ρ
L
gσ
_1
4
and β must always be ≥ 0. Therefore if a negative value
of β is obtained, β=0.
Once the actual liquid holdup E
L
(θ) is calculated, the mixture density ρ
m
is ob
tained. Mixture density, in turn, is used to calculate the pressure change due to the
hydrostatic head of the vertical component of the pipe
∆P
HH
=
ρ
m
gLsin(θ)
144g
c
(3.32)
3.3.1.3 Frictional Pressure Loss
In order to calculate frictional losses, a normalizing friction factor f
NS
is used. To
determine f
NS
, we utilize the Moody Friction factor calculated using the Colebrook
equation. For this purpose, the noslip Reynolds number is used:
Re
NS
=
ρ
NS
V
m
D
µ
NS
(3.33)
Based on experimental data, Beggs and Brill presented a correlation for the ratio of
the twophase friction factor f
tp
and the normalizing (noslip) friction factor resulting
in the following exponential equation:
f
tp
= f
NS
e
S
(3.34)
The value of S depends on the noslip and the actual liquid holdup:
S =
ln(Y )
−0.0523 + 3.182(ln(Y )) −0.8725(ln(Y ))
2
+ 0.01853(ln(Y ))
4
(3.35)
where Y =
C
L
E
L
(θ)
2
3.3. TWO PHASE FLOW CORRELATIONS 39
1: if Y ≥ 1 and Y ≤ 1.2 then
2: S ←ln(2.2Y −1.2)
3: else
4: S ←
ln(Y )
−0.0523+3.182(ln(Y ))−0.8725(ln(Y ))
2
+0.01853(ln(Y ))
4
5: end if
Finally, the expression for pressure loss due to friction is:
∆P
f
=
2f
tp
V
2
m
ρ
NS
L
144g
c
(3.36)
3.3.2 Aziz, Govier and Fogarasi Model
Fontanilla[19] investigated applicability of the Aziz et al. [29] mechanistically based
pressure drop correlation for downwards steam ﬂow. As Fontanilla stated for Aziz et
al. [29], this correlation that was strictly developed for upward vertical ﬂow needs
some modiﬁcation in order to apply to downward ﬂow. Those modiﬁcations are
included in the bubble ﬂow and in the slug ﬂow patterns. Due to the large volume
of steam(gas) as compared to the water that is present in most steam injection wells,
the annular mist ﬂow pattern is very common.
3.3.2.1 Flow Pattern Determination
In Figure 3.14 we can see the ﬂowpattern map for diﬀerent ﬂow types.
N
x
= V
Sg
_
ρ
g
0.0764
_1
3
__
72
σ
L
_
_
ρ
L
62.4
_
_1
4
(3.37)
N
y
= V
SL
__
72
σ
L
_
_
ρ
L
62.4
_
_1
4
(3.38)
40 CHAPTER 3. MODEL FORMULATION
Figure 3.14: Flow Pattern map for Aziz et al. (retrieved from[14]).
In the Figure 3.14 we have several lines that represents ﬂowpattern transitions
and they are deﬁned as:
N
1
= 0.51(100N
y
)
0.172
(3.39)
N
2
= 8.6 + 3.8N
y
(3.40)
N
3
= 70(100N
y
)
−0.152
(3.41)
where v
SL
is in ft per second, v
S
is in feet per second, ρ
g
is in pounds per cubic feet,
ρ
L
is in pounds per cubic feet and σ
L
is in dynes per centimeter.
Bubble Flow
The bubble ﬂow pattern is characterized by small bubbles of steam(gas) dispersed in
a continuous water phase. In downward ﬂow, the diﬀerence in densities of the two
phases causes the bubbles to travel at a velocity lesser than the average velocity of the
mixture. Bubble ﬂow exists if N
x
< N
1
. Liquid holdup for bubble ﬂow is calculated
from
H
L
= 1 −
v
Sg
v
bf
(3.42)
where v
bf
is the rise velocity of small gas bubbles in a ﬂowing liquid. This velocity is
3.3. TWO PHASE FLOW CORRELATIONS 41
predicted for downward ﬂow as follow
v
bf
= 1.2v
m
−v
bs
(3.43)
where the ﬁrst term is the approximate velocity of the ﬂuid mixture, accounting for
the nonuniform velocity and bubble concentration proﬁles across the cross section,
and v
bs
is the rise velocity of a continuous swarm of small bubbles in a static liquid
column. The v
bs
term is predicted from
v
bs
= 1.41
_
σ
L
g(ρ
L
−ρ
g
)
ρ
2
L
_1
4
(3.44)
The frictional component of the pressure gradient is determined as
∂p
∂Z
=
fρ
s
v
2
m
2d
(3.45)
where ρ
s
is determined from equation (4.34) and friction factor from Moody Friction
factor calculation using the Colebrook equation[27].
ρ
s
= ρ
L
H
L
+ρ
g
(1 −H
L
) (3.46)
N
Re
=
ρ
L
v
m
d
µ
L
(3.47)
The acceleration component of the pressure gradient is considered to be negligible for
bubble ﬂow.
42 CHAPTER 3. MODEL FORMULATION
Slug Flow
Slug ﬂow exists if N
1
< N
x
< N
2
for N
y
< 4 or N
1
< N
x
< 26.5 for N
y
≥ 4.
The liquid holdup for slug ﬂow is also calculated from Eq.(4.30) and Eq.(4.37). For
slug ﬂow, however, the bubblerise velocity in a static liquid column is based on large
bubble. Aziz et al.[29] state that
v
bs
= C
¸
gd(ρ
L
−ρ
g
)
ρ
L
(3.48)
where C was given by Wallis as
C = 0.345
_
1 −e
(−0.029Nv)
¸
_
1 −e
(
3.37−N
E
m
)
_
(3.49)
and
N
E
=
gd
2
(ρ
L
−ρ
g
)
σ
L
(3.50)
N
v
=
_
d
3
gρ
L
(ρ
L
−ρ
g
)
µ
L
(3.51)
and m is determined as
1: if N
v
≥ 250 then
2: m ←10
3: else
4: if 250 > N
v
> 18 then
5: m ←69N
−0.35
v
6: end if
7: if N
v
≤ 25 then
8: m ←25
9: end if
10: end if
3.3. TWO PHASE FLOW CORRELATIONS 43
The friction pressuregradient component for slug ﬂow is determined from
_
dp
dZ
_
f
=
fρ
L
H
L
v
2
m
2d
(3.52)
The friction factor is obtained from a Moody friction factor and the Colebrook
equation[27]. The Reynolds number is given as
N
Re
=
fρ
L
v
m
d
µ
L
(3.53)
The acceleration pressuregradient component was considered negligible for slug ﬂow.
Mist Flow
Mist ﬂow exists when N
x
> N
3
for N
y
< 4 or N
x
> 26.5 for N
y
> 4. Aziz et al.[29]
recommended the Duns and Ros [28] mistﬂow method be used to calculate pressure
gradient for this ﬂow pattern.
Transition Flow
The transition region exists when N
2
< N
x
< N
3
for N
y
< 4. In Figure 4.8, the tran
sition region does not exist for N
y
> 4. When the transition region is predicted, the
pressure gradients must be calculated with both the slugﬂow and mistﬂow equations.
To obtain the pressure gradient, linear interpolation is performed.
dp
dZ
= A
_
dp
dZ
_
slug
+ (1 −A)
_
dp
dZ
_
mist
(3.54)
where A =
N
3
−Nx
N
3
−N
2
3.3.2.2 Modiﬁcations
AlNajjar and AlSoof [5] showed that improved results could be obtained with Aziz
et al. method if the ﬂow pattern map in Figure 3.14 was replaced with the Duns and
Ros map. Their conclusion was based on a comparison of the predicted and measured
44 CHAPTER 3. MODEL FORMULATION
pressure drops for 80 tests on 15 ﬂowing wells in Iraq.
Chapter 4
Eﬀect of NonCondensable Gas
(N
2
)
In our calculations we used N
2
as the noncondensable gas. We explore the partial
pressure eﬀect of N
2
on downhole steam quality, temperature, pressure and as well
as heat loss during the steam injection. Addition of noncondensable gas into our
system makes our calculation a little complex [10]. The pressure of the vapor phase
is now a sum of steam partial pressure and gas partial pressure:
P
total
= P
steam
+ P
N
2
(5.1)
where P
total
is the total pressure of the gas phase, P
steam
is the partial pressure of the
steam and P
N
2
is the partial pressure of the N
2
. We assume N
2
only exists in the
vapor phase with steam and has the same temperature as the steam. N
2
is assumed
not to carry heat, so contribution to the mixture enthalpy is zero. N
2
, however,
provides partial pressure during steam injection. Neglecting the enthalpy of N
2
is a
good approximation because steam enthalpy is several orders of magnitude larger in
comparison to N
2
.
45
46 CHAPTER 4. EFFECT OF NONCONDENSABLE GAS (N
2
)
The total mol fraction of the liquid and gas phase is
x
i
= 1.0 (5.2)
y
i
= 1.0 (5.3)
where x is the liquid mole fraction of the components, y is the vapor mole fraction of
the components. Because we have only liquid water in our system inequlibrium with
steam and N
2
,
x
i
= 1.0 is always equal to x
water
.
Calculation of the partial pressure of the gas components is
P
N
2
= y
N
2
∗ P
total
(5.4)
P
steam
= y
steam
∗ P
total
(5.5)
Suppose we have a system with given injection rate, steam quality, steam tem
perature or pressure (one determines the other), time and mol fraction of N
2
. So,
mole fraction of steam is found from Eq. 5.3 and
y
2
= y
N
2
+ y
steam
= 1.0 yields
that y
steam
= 1 − y
N
2
. With a given undilited steam pressure or temperature (one
determines other), we ﬁnd the total pressure as
P
total
=
P
steam
y
steam
=
P
steam
1 −y
N
2
(5.6)
Once we ﬁnd total pressure we continue to ﬁnd other input parameters in our
calculation. Total mass of the steam is found as
W
mX
= Wm∗ t ∗ x (5.7)
Having the total mass of the steam in our system, we ﬁnd the total moles of the steam
47
as
Mole
X
=
W
mX
∗ 454
18.02
(5.8)
With known steam mole fraction, we ﬁnd how many moles of N
2
is in the system
as
y
steam
=
Mole
X
Mole
X
+Mole
N
2
(5.9)
Because y
steam
is known and y
N
2
is known, there is only one unknown in Eq. 5.9.
Mole
N
2
is easily found using Eq. 5.9. The reason we calculated mole of the N
2
is
because it is not changed entire process. Meaning that mass of the N
2
is constant al
the time. The only thing is going to change in each interval is that mol fraction of the
N
2
, and it is going to increase due to the condensation of steam. Another important
assumption is that the viscosity of mixture of N
2
and steam is the as same as the
viscosity of the undiluted steam. The equation for viscosity is given by Eq.3.23.
We assume that in two phase ﬂow, cases occur where the gas partial pressure is
signiﬁcant compared to the steam partial pressure. When this is the case the gas and
steam act as perfect gases. Thus we use ideal gas law
PV = n ∗ R ∗ T (5.10)
where P is the total pressure of the system in kilo pascal, n is the mole fraction of the
component, R is the universal gas constant 8.314 J/Kmol and T is the temperature
in Kelvin, K =
o
C + 273.15.
Total volume of the steam and N
2
is obtained using ideal gas law as
V
steam
=
Mole
X
∗ R ∗ T
steam
P
steam@initial
(5.7)
V
N
2
=
Mole
N
2
∗ R ∗ T
steam
P
steam@initial
(5.8)
48 CHAPTER 4. EFFECT OF NONCONDENSABLE GAS (N
2
)
Once steam volume and N
2
volume are obtained, it is easy to ﬁnd density of the total
gas mixture. Another, potentially more accurate, method is to obtain the steam
molar volume from the steam tables. In either case, the overall volume is found by
applying Amagat’s law. That is, the volumes of each component add ideally and
there is no volume change upon mixing. The density of the gas mixture and liquid
mixture are used in our calculation. In each segment of the wellbore, this calculation
has to be repeated until steam reaches the reservoir.
Chapter 5
Graphical User Interface (GUI)
Our GUI provides several inputs parameters and options that are useful in the cal
culations. The user can choose diﬀerent twophase correlations and get the results
as well as choose previous calculation results for postprocessing. The theory behind
calcualtions is well described in preceding chapters and the Appendix. The overall
heat transfer coeﬃcient is implemented both following Willhite’s approach[49] and
Ramey’s time function[42]. This is similar to Fontanilla [19]. This chapter shows how
we get the results using our GUI.
GUI for the Main program
The GUI is demonstrated with snapshot of the program. The ﬁrst snapshot gives the
background for the program. The second snapshot is about calculations for onshore
environment with insulated and without insulated tubing. Also, we provide solutions
for the addition of noncondensable gas N
2
for both onshore and oﬀshore environ
ments. The third snapshot is the input parameters for the oﬀshore environment. The
fourth snapshot is the output tables both onshore and oﬀshore environments. The
ﬁfth snapshot is about the postprocessing for viewing the results in ﬁgures for steam
quality, steam pressure, steam temperature and heat loss along the wellbore. All
49
50 CHAPTER 5. GRAPHICAL USER INTERFACE (GUI)
calculations take into account twophase ﬂow correlations.
First snapshot of the program is shown Figure 5.1, it is seen that there are several
input parameters for speciﬁcally steam injection parameters for onshore cases. User
can choose diﬀerent insulation materials and twophase ﬂow correlations. Figure 5.2
is the similar input parameters except riser radius and sea temperature inputs. User
also able to choose insulation materials and two phase ﬂow correlation to calculate
steam temperature, steam pressure, steam quality and heat loss values. In third
ﬁgure that is Figure 5.3, after the calculation is done either for onshore or oﬀshore
cases, results are shown this part and user able to save those datas to excel ﬁles or
clear them. Last one is the Figure 5.4, is postprocessing the data that obtained from
calculations either onshore or oﬀshore is designed speciﬁcally to visualize the results
and to save the ﬁgures or delete it.
Figure 5.1: User interface developed GUI for onshore calculations.
51
Figure 5.2: User interface developed GUI for oﬀshore calculations.
Figure 5.3: User interface developed GUI for both onshore and oﬀshore results.
52 CHAPTER 5. GRAPHICAL USER INTERFACE (GUI)
Figure 5.4: User interface developed GUI postprocessing for both onshore and oﬀ
shore results.
Chapter 6
Results and Comparisons
In this part of the study, surface line, sea and sea ﬂoor to reservoir heat loss calcu
lations are presented using limited data provided by Prats [41]. We compared our
results with Prats’s [41] results. It is shown that our results match with Prats’s [41]
results for surface lines. For injection wells, our results are a little higher than his
results. With this we found his minor mistake for onshore injection well and reported
here. And then results are extended to 2 ﬁeld cases [11]. Moreover, noncondensable
gas addition calculations are done with diﬀerent injection rate, steam quality, steam
temperature, diﬀerent depth and diﬀerent percentage of N
2
addition in order to see
the sensitivity of steam temperature, steam pressure, steam quality and heat loss.
6.1 Examples for heat loss calculation
Examples with known solutions are presented for surface line and onshore calcualtions
with insulated or without insulated tubing to verify the new code. There is no example
for oﬀshore heat loss calculation, so we take input values from Prats [41] and modify
the calculations to illustrate the hole of insulation. This part of results are only
considered for heat loss aspects.
53
54 CHAPTER 6. RESULTS AND COMPARISONS
6.1.1 Example 10.1 from Prats
Steam at 550
o
F is injected through 4in. N80 pipe at a rate of 229 B/D. Prats [41]
asks ”ﬁnd the steady state heat loss per year per 100 ft of pipe when the pipe is (1) is
insulated with 3 in of calcium silicate and (2) not insulated  i.e., bare.” The average
yearly temperature is 60
o
F, and the prevailing winds have an average velocity of 20
mph normal to the injection line. The input parameters are listed in Table 6.1.
h
fc
=
18v
0.6
w
r
0.6
ins
r
ins
(6.1)
With insulation, the over all speciﬁc thermal resistance is calculated from Eq. 3.2
and plotted into Figure 6.1 and for without insulation plotted into Figure 6.2.
Accordingly, the amount of heat lost from 100ft length of pipe over a period of
1 year is Q
l
=1.1581*10
8
BTU for the calcium silicate (highest heat loss), for black
aerogel heat loss is Q
l
=2.0302*10
7
BTU (lowest heat loss).
Without insulation and at a surface temperature near 550
o
F, radiation heat losses
are important. The sum of the coeﬃcients of heat transfer due to the radiation and
free(or natural) convection for a horizontal pipe is given Table 3.1 for a several pipe
sizes and temperatures. Eq. 3.2 is also applied here in order ﬁnd the heat loss for
without insulation case.
Thus, the amount of heat lost from a 100ft length of pipe over a period of 1 year
is: Q
l
=6.095*10
9
BTU when the pipe is bare. Therefore, the insulation reduces heat
losses by a factor of about 50 when using calcium silicate, however if you use aerogel
it would be 280. Because one barrel of oil is roughly 6.0*10
6
BTU, the reduction in
yearly heat losses resulting from insulating the pipe amounts to more than 1000 bbl
of fuel for a 100ft length of pipe. When you use calcium silicate you would consume
100 bbl, on the other side when you use aerogel you would consume 16 bbl for 100
ft length surface pipe. When we think about deeper wells using insulation material
6.1. EXAMPLES FOR HEAT LOSS CALCULATION 55
should be considered especially aerogel. Table 6.1 is used when we calculate the heat
losses for the uninsulated case.
In Example 10.1 with insulations, Prats [41] used calcium silicate as the insulation
material and get the heat loss as, Q
l
=1.16*10
8
BTU, over the year. Our result
for using calcium silicate is Q
l
=1.1581*10
8
BTU and it is consistent with Prats’s
example as shown in Figure 6.1. Without insulation, Prats [41] stated that the
coeﬃcient of heat transfer due to radiation and forced convection is estimated to be
330 BTU/ft
2
−D−
o
F, however, he used 110 BTU/ft
2
−D−
o
F and get the heat loss
results over the one year period is Q
l
=6.21*10
9
BTU. In our calculation we ﬁnd this
minor mistake from Prats calculation, and it causes little deviation from the exact
results. This shows that the real heat loss from bare tubing is Q
l
=6.095*10
9
BTU as
shown in Figure 6.2. Still those results match well.
In Table 6.1, there are several column headings. Abbreviations are explained as
”SLwithIns” is the surface line with insulated tubing, ”SLwithoutIns” is the surface
line without insulated tubing, ”SeawithIns” is the oﬀshore case with insulated tubing,
”SeawitouthIns” is the oﬀshore case without insulated tubing, ”2ResWithIns” is the
sea ﬂoor to reservoir case with insulated tubing, and ”2ResWithoutIns” is the sea
ﬂoor to reservoir case without insulated tubing.
5
6
C
H
A
P
T
E
R
6
.
R
E
S
U
L
T
S
A
N
D
C
O
M
P
A
R
I
S
O
N
S
Table 6.1: Input parameters from Prats [41] as used for diﬀerent example calculations.
Parameters SLwithIns SLwithoutIns SeaWithIns SeaWithoutIns 2ResWithIns 2ResWithoutIns
T
steam
550 550 600 600 600 600
T
average
60 60 70 70 100 100
r
ti
0.1478 0.1478 0.14 0.14 0.14 0.14
r
to
0.1667 0.1667 0.1458 0.1458 0.1458 0.1458
r
ins
0.41467 No Insulation 0.2292 No Insulation 0.2292 No Insulation
r
riserin
0 0 0.6 0.6 0 0
r
riserout
0 0 0.75 0.75 0 0
h
f
2000 2000 0 0 0 0
h
pi
∞ ∞ 0 0 0 0
h
po
2000 2000 0 0 0 0
λ
pipe
25 25 25 25 0 0
λ
cem
0 0 24 24 24 24
λ
ins
0.166/24 No Insulation 0.166/24 No Insulation 0.166/24 No Insulation
time 365*24 365*24 21*24 21*24 21*24 21*24
PipeLength 100 100 164 164 1000 1000
riserin
0 0 0.9 0.9 0 0
riserout
0 0 0.9 0.9 0 0
rto
0 0 0.9 0.9 0 0
riserin
0 0 0 0 0.9 0.9
riserout
0 0 0 0 0.9 0.9
rto
0 0 0 0 0.9 0.9
6
.
1
.
E
X
A
M
P
L
E
S
F
O
R
H
E
A
T
L
O
S
S
C
A
L
C
U
L
A
T
I
O
N
5
7
Table 6.2: Radiationnatural convection coeﬃcient of heat transfer.
Diameter 130 180 230 280 330 380 480 580 680 780 880 980 1080 1180 1280
0.50 50.9 59.5 66.2 74.4 81.8 90.0 107.0 127 149 174 202 234 269 307 352
1.00 48.7 57.1 63.6 71.5 79.0 86.9 104.0 124 146 171 198 230 265 304 348
2.00 46.3 54.5 60.5 68.4 75.4 83.3 100.0 120 141 166 194 225 260 299 343
4.00 44.2 51.8 57.8 65.3 72.2 79.9 96.5 116 137 162 189 221 256 294 338
8.00 42.2 49.4 70.1 62.4 69.4 76.8 93.1 112 134 158 186 217 252 290 334
12.00 41.0 48.2 53.8 61.0 67.7 75.1 91.9 111 132 156 184 215 250 289 332
24.00 39.4 46.3 51.6 58.8 65.3 72.7 88.8 108 129 153 180 212 247 286 329
58 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.1: Surface lines heat loss calculation with six diﬀerent insulation materials.
Figure 6.2: Surface Heat Loss calculation without insulation.
6.1. EXAMPLES FOR HEAT LOSS CALCULATION 59
6.1.2 Example for Oﬀshore
Figure 6.3: Heat loss from sea level to sea ﬂoor with six diﬀerent insulations.
Figure 6.4: Heat loss from sea level to sea ﬂoor without insulation.
60 CHAPTER 6. RESULTS AND COMPARISONS
Under 600
o
F, steam injection for oﬀshore heat loss is calculated with insulated
material and uninsulated cases. Sea temperature is taken as 70
o
F for 164 ft water
depth. This is relatively shallow. Heat loss estimation for 21 days of injection shows
that the greatest the thermal conductivity values the greatest the heat losses observed.
For instance using calcium silicate (2.87 ∗ 10
4
) BTU gives four times more heat loss
than black aerogel (0.75 ∗ 10
4
) BTU. In addition, the bare tubing (2.1 ∗ 10
5
) BTU
case also gives about 8 times more heat loss than using lowest thermal conductivity
insulation materials. It is clear that for 21 days period heat loss are substantial for
uninsulated wells.
6.1.3 Example 10.2 from Prats
Steam at 600
o
F is injected down 3.5in. tubing set on a packer in 9 5/8 in., 53.5
lbm/ft N80 casing. The annulus contains a stagnant gas at zero gage pressure at
wellhead, and the casing is cemented to surface in a 12in. hole. The tubing is
insulated with 1 in. of calcium silicate, the insulation being held in place and sealed
from accidental entry of liquids in the annulus by a very thin sheath of aluminum. A
temperature survey in the well indicates a mean surface temperature of 100
o
F over
the 1000ft depth. Estimate the rate of heat loss 21 days after steam injection started,
as well as the casing temperature. There is no altered zone near the boreholes. Input
parameters are taken from Table 6.1.
Heat transfer across the gas ﬁlled annulus is by radiation and natural convection.
Radiation is sensitive to the temperature levels and emissivities (∈) of the surfaces.
The temperature at the surface of the insulation (T
ins
) and that at the inner radius
of the casing (T
ci
), together with the emissivities at these surfaces (∈
ins
) and (∈
ci
),
aﬀect the radiation heat losses across the annular space between the insulated casing.
The calculation procedure is well explained in the Prats book[41].
Prats [41] assumed that for shallow reservoir temperature change does not vary
6.1. EXAMPLES FOR HEAT LOSS CALCULATION 61
much with depth such as 1000 ft. He also stated that when the steam injection hap
pens, shallow reservoir temperature stays constant. This is why using a single value
of R
h
may provide a close enough estimate of heat losses and wellbore temperatures.
This is good approach in terms of calculation and simplicity.
Example 10.2 from Prats [41] got heat loss Q
l
=4.46*10
6
BTU for 21 days that
corresponds to an equivalent energy content of 6.0*10
6
BTU/bbl of fuel, the daily
heat loss from such an insulated well corresponds to less than 1 bbl of fuel. In our
calculation we get Q
l
=4.6*10
6
BTU that is slightly diﬀerent than Prats results that is
shown in Figure 6.5. It is because of several input parameters that diﬀers somewhat.
Discussion of heat loss so far, considers the entire well as a unit. It is because of
assuming temperature of the ﬂuid does not change with depth for shallow reservoirs.
In our program validation we do not assume this and we take into account temperature
change along the wellbore.
62 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.5: Heat loss calculation using diﬀerent insulation materials based on Example
10.2 from Prats[41].
Figure 6.6: Heat loss calculation without using insulation materials based on Example
10.2 from Prats [41].
6.2. PROGRAM VALIDATION 63
6.2 Program Validation
Table 6.3: Field data parameters for ﬁeld data 1 and ﬁeld data 2 [19].
Input Parameters Field Data 1 Field Data 2
D
ti
0.177 0.177
r
to
0.104166667 0.104166667
r
ins
No Insulation No Insulation
r
ci
0.166666667 0.166666667
r
co
0.1875 0.1875
r
h
0.6 0.6
k
e
1 1
α
E
0.0286 0.0286
k
cem
0.2 0.2
to
0.9 0.9
ci
0.9 0.9
EARTH
0.94 0.94
W
m
4640 4850
x 0.8 0.8
p
wh
250 250
T
m
50 50
Depth 1600 1600
P
an
14.7 14.7
t 71 117
There are 2 ﬁeld cases where steam temperature is reported in the literature [11].
The Martha Bigpond well data is retrieved from the paper. Fontanilla and Aziz [20]
64 CHAPTER 6. RESULTS AND COMPARISONS
referred to the same ﬁeld data from Bleakley’s paper [11] as ﬁeld data 1 and 2. These
cases are to referred in our calculations as ﬁeld data 1 and ﬁeld data 2 as well. Input
parameters are tabulated in Table 6.3. Besides Beggs and Brill [13, 14] and Aziz et
al. [29], Hasan and Kabir model [23, 40] and the Drift Flux model [21] are applied in
our calculations.
After applying the modiﬁed Fontanilla[20] approach with modiﬁed correlations,
we got promising results using the Beggs and Brill [13, 14] approach for multiphase
ﬂow with ﬁeld data 1. The other three correlations also get good agreement with
Fontanilla’s results as shown in Figure 6.7 and 6.8. When we look at the steam quality
versus depth, the Beggs and Brill over predicts the quality values that Fontanilla got
for ﬁeld data 1 in Figure 6.9. The other three correlations that we implemented have
almost the same results. It may be because of applying diﬀerent correlations to get
saturated steam properties during calculations.
In Figure 6.10, we can see the heat loss during steam injection is almost same
values for the four diﬀerent correlations of multiphase ﬂow.
6.2. PROGRAM VALIDATION 65
355 360 365 370 375 380 385 390 395 400 405
0
200
400
600
800
1000
1200
1400
1600
Steam Temperature in
o
F
D
e
p
t
h
(
f
t
)
Field Data 1
Fontanilla
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
371
o
F 364
o
F
358
o
F 368
o
F
Figure 6.7: Comparison of steam temperature with ﬁeld data 1 and twophase corre
lations.
140 160 180 200 220 240
0
200
400
600
800
1000
1200
1400
1600
Steam Pressure in psi
D
e
p
t
h
(
f
t
)
Field Data 1
Fontanilla
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
150
160
166
168
Figure 6.8: Comparison of steam pressure with ﬁeld data 1 and twophase correlations.
66 CHAPTER 6. RESULTS AND COMPARISONS
0.7 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 0.8
0
200
400
600
800
1000
1200
1400
1600
Steam Quality
D
e
p
t
h
(
f
t
)
Fontanilla
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
0.745
0.703
Figure 6.9: Calculated steam quality with diﬀerent twophase correlations based on
ﬁeld data 1.
0 1 2 3 4 5 6 7 8 9 10
x 10
5
0
200
400
600
800
1000
1200
1400
1600
Heat Loss in BTU
D
e
p
t
h
(
f
t
)
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
Figure 6.10: Calculated heat loss calculation with insulated tubing based on ﬁeld
data 1.
6.2. PROGRAM VALIDATION 67
Figure 6.11 and 6.12 also show results after we run our simulator for ﬁeld test
data 2. Again, we can see that the modiﬁed Beggs and Brill method gives a bet
ter result than the Fontanillla approach compared to the ﬁeld data 1. The other
three correlations give results similar to Fontanilla’s result. Although, we got good
results for steam temperature and pressure values for Beggs and Brill, our calculation
shows a little overprediction of values for steam quality as shown on Figure 6.13 from
Fontanilla’s result.
355 360 365 370 375 380 385 390 395 400 405
0
200
400
600
800
1000
1200
1400
1600
Steam Temperature in
o
F
D
e
p
t
h
(
f
t
)
Field Data 2
Fontanilla
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
Figure 6.11: Comparison of steam temperature with ﬁeld data 2 and twophase cor
relations.
68 CHAPTER 6. RESULTS AND COMPARISONS
140 150 160 170 180 190 200 210 220 230 240 250
0
200
400
600
800
1000
1200
1400
1600
Steam Pressure in psi
D
e
p
t
h
(
f
t
)
Field Data 2
Fontanilla
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
Figure 6.12: Comparison of steam pressure with ﬁeld data 2 and twophase correla
tions.
0.68 0.7 0.72 0.74 0.76 0.78 0.8
0
200
400
600
800
1000
1200
1400
1600
Steam Quality
D
e
p
t
h
(
f
t
)
Fontanilla
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
Figure 6.13: Calculated steam quality with diﬀerent twophase correlations based on
ﬁeld data 2.
6.2. PROGRAM VALIDATION 69
0 2 4 6 8 10 12 14 16
x 10
5
0
200
400
600
800
1000
1200
1400
1600
Heat Loss in BTU
D
e
p
t
h
(
f
t
)
Aziz, Govier and Fogarasi
Beggs & Brill
Hasan & Kabir
DriftFlux
Figure 6.14: Calculated heat loss calculation with insulated tubing based on ﬁeld
data 2.
340 350 360 370 380 390 400 410
0
200
400
600
800
1000
1200
1400
1600
Steam Temperature in
o
F
D
e
p
t
h
(
f
t
)
Field Data 1
Our Model(with Aziz)
Our Model(with Beggs&Brill)
Fontanilla Model(with Aziz)
Fontanilla Model(with Beggs&Brill)
Figure 6.15: Comparison of our model with Fontanilla’s model steam temperature
with ﬁeld data 1.
70 CHAPTER 6. RESULTS AND COMPARISONS
120 140 160 180 200 220 240 260
0
200
400
600
800
1000
1200
1400
1600
Steam Pressure in psi
D
e
p
t
h
(
f
t
)
Field Data 1
Our Model(with Aziz)
Our Model(with Beggs&Brill)
Fontanilla Model(with Aziz)
Fontanilla Model(with Beggs&Brill)
Figure 6.16: Comparison of our model with Fontanilla’s model steam pressure with
ﬁeld data 1.
0.7 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 0.8
0
200
400
600
800
1000
1200
1400
1600
Steam quality
D
e
p
t
h
(
f
t
)
Our Model(with Aziz)
Our Model(with Beggs&Brill)
Fontanilla Model(with Aziz)
Fontanilla Model(with Beggs&Brill)
Figure 6.17: Comparison of our model with Fontanilla’s model steam quality with
ﬁeld data 1.
When we look at the results that we got for both ﬁeld data 1 and 2 cases, we had
an opportunity to compare the results with Fontanilla’s approach. The ﬁgures above
6.2. PROGRAM VALIDATION 71
for the ﬁeld data 1, and our implementation of Beggs and Brill correlations starts
converging to the real data when we reach the bottom of the wellbore in Figure 6.15.
On the other hand Fontanilla’s Beggs and Brill implementation has less accurate
results than ours. Our model using Aziz et al. has good agreement with Fontanilla’s
Beggs and Brill results. Looking at the results, Fontanilla’s Aziz et al. calculations
give almost perfect results with the ﬁeld data until 1000 ft. Reuslts then deviate from
the real values and go out of the range. In Figure 6.16, we can also see the same trend
with Figure 6.15. It comes to check steam quality, although we are able to calculate
steam quality with given injection rate and input parameters, in the literature we
could not ﬁnd the steam quality data. Therefore we only compare the results we have
in our model and Fontanilla’s model, in Figure 6.17. We can conclude the comparison
saying that the Aziz et al. model starts converging from the beginning and obtains
similar values. On the other hand, Beggs and Brill results diﬀer quite a bit between
the two methods.
330 340 350 360 370 380 390 400 410
0
200
400
600
800
1000
1200
1400
1600
Steam Temperature in
o
F
D
e
p
t
h
(
f
t
)
Field Data 2
Our Model(with Aziz)
Our Model(with Beggs&Brill)
Fontanilla Model(with Aziz)
Fontanilla Model(with Beggs&Brill)
Figure 6.18: Comparison of our model with Fontanilla’s model steam temperature
with ﬁeld data 2.
72 CHAPTER 6. RESULTS AND COMPARISONS
100 120 140 160 180 200 220 240 260
0
200
400
600
800
1000
1200
1400
1600
Steam Pressure in psi
D
e
p
t
h
(
f
t
)
Field Data 2
Our Model(with Aziz)
Our Model(with Beggs&Brill)
Fontanilla Model(with Aziz)
Fontanilla Model(with Beggs&Brill)
Figure 6.19: Comparison of our model with Fontanilla’s model steam pressure with
ﬁeld data 2.
0.68 0.7 0.72 0.74 0.76 0.78 0.8
0
200
400
600
800
1000
1200
1400
1600
Steam quality
D
e
p
t
h
(
f
t
)
Our Model(with Aziz)
Our Model(with Beggs&Brill)
Fontanilla Model(with Aziz)
Fontanilla Model(with Beggs&Brill)
Figure 6.20: Comparison of our model with Fontanilla’s model steam quality with
ﬁeld data 2.
In the ﬁeld data 2, we got the results and plot them as shown in Figures 6.18  6.20.
In Figure 6.18 our model with Beggs and Brill correlations gives very good agreement
6.3. ONSHORE ENVIRONMENTS 73
with the real data and converges to the exact value. In contrast, Fontanilla’s method
with the Beggs and Brill approach gives less accurate results than our model. In
our model, the Aziz et al. method, results better match ﬁeld data as compared
to Fontanilla’s Aziz et al. approach, however, Fontanilla’s Aziz et al. approach
initially gives great results and captures almost exact values for several hundreds ft
then deviates substantially from the reality. We can also conclude with saying that
we improved Fontanilla’s approach using several twophase correaltions. This trends
continue in Figure 6.19 with pressure values. The last comparison is steam quality.
We can say that our model with Beggs and Brill twophase ﬂow correlations has less
quality drop than Fontanilla’s Begss and Brill model in Figure 6.20. The Aziz et al.
approach has similar steam quality values in both models.
6.3 Onshore environments
For both onshore and oﬀshore environments with insulated or uninsulated tubing
input parameters are used for ﬁeld data 1. The surrounding temperature, however, is
taken as 122
o
F, time is 1 year and injection rate 4850 lbm/hr similar to the onshore,
oﬀshore and noncondensable gas cases. For the oﬀshore part, ﬁeld data 1 values are
used as an input and exception is also used here with additional input for oﬀshore
sea temperature is taken as 70
o
F.
Results in Figures 6.21 and 6.22 are insulated by black aerogel, and various two
phase correlations are used. Three diﬀerent steam injection temperatures and two
diﬀerent depths are considered for the onshore environment. Steam temperature,
steam pressure, steam quality and heat loss values are investigated. When the steam
temperature is 400
o
F, the Beggs and Brill twophase ﬂow correlation model’s tem
perature proﬁle or pressure proﬁle drops dramatically. On the other hand, three
other twophase correlations converge to the same values both on temperature proﬁle
74 CHAPTER 6. RESULTS AND COMPARISONS
and pressure proﬁle. When we increase injection temperature to 500
o
F, the Beggs
and Brill twophase ﬂow correlation model’s temperature or pressure proﬁle drops
less than the other correlations until 7500 ft then converge to the others. When
the depth is greater than 8000 ft, it gives a little greater value than other correla
tions. Interestingly, the behavior of the Beggs and Brill model when the injection
temperature increased to 600
o
F gives less temperature drop than all the other three
correlations. Figures 6.25 and 6.26 show that with diﬀerent temperature and depth
all the ﬂow regimes have similar values in steam temperature, steam pressure, and
steam quality comparing with insulated case. The diﬀerences between insulated and
uninsulated tubing is the heat loss and steam quality values. Figure 6.23 and Figure
6.27 show results. The steam quality values with insulated tubing when compared to
uninsulated tubing cases diﬀer signiﬁcantly. Without insulation, there is greater heat
loss and quality decreases to a greater extent as compared to the insulated case. It is
also observed that, the Beggs and Brill model, with increasing injection temperature
overpredicts the steam quality values. Figures 6.24 and 6.28 give the information that
increasing injection steam temperature causes greater heat loss when comparing 500
o
F and 600
o
F cases.
Similar trends are obtained during the steam injection in an oﬀshore environment
as well. Although the oﬀshore depth of 200 ft is relatively shallow and greater pres
sure drop and temperaure drop for small interval does not really eﬀect the behavior
trend of the steam properties at downhole conditions. The greater the steam tem
perature, the greater steam quality and steam properties are obtained in both cases.
The greatest heat loss is obtained using calcium silicate which has greatest thermal
conductivity value among the six diﬀerent insulations. Figures presented in Appendix
C also have similar behavior in our example, therefore this explanation will help to
understand those ﬁgures as well.
6.3. ONSHORE ENVIRONMENTS 75
6.3.1 Examples with Insulation Materials
UsingBlackAerogelλ
BA
= 0.0069BTU/(ft −hr −
o
F)
Figure 6.21: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using black
aerogel.
Figure 6.22: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using black aerogel.
76 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.23: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using black aerogel.
Figure 6.24: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with using black aerogel.
6.3. ONSHORE ENVIRONMENTS 77
6.3.2 Examples without Insulation Materials
Figure 6.25: Steam temperature distribution , 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
Figure 6.26: Steam pressure distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
78 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.27: Steam quality distribution, 1 year, T
m
= 122
o
F and injection rate 4850
lbm/hr without insulation.
Figure 6.28: Heat loss distribution, 1 year, T
m
= 122
o
F and injection rate 4850
lbm/hr without insulation.
6.4. OFFSHORE ENVIRONMENTS 79
6.4 Oﬀshore Environments
6.4.1 Examples with Insulation Materials
UsingBlackAerogelλ
BA
= 0.0069BTU/(ft −hr −
o
F)
Figure 6.29: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black aerogel.
Figure 6.30: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black aerogel.
80 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.31: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black aerogel.
Figure 6.32: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr for black aerogel.
6.4. OFFSHORE ENVIRONMENTS 81
6.4.2 Examples without Insulation Materials
Figure 6.33: Steam temperature distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
Figure 6.34: Steam pressure distribution, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
82 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.35: Steam quality distribution, 1 year, T
m
= 122
o
F and injection rate 4850
lbm/hr without insulation.
Figure 6.36: Heat loss distribution, 1 year, T
m
= 122
o
F and injection rate 4850
lbm/hr without insulation.
6.4. OFFSHORE ENVIRONMENTS 83
Figure 6.37: Pressure drop distribution and formation pressure (green dots), 1 year,
T
m
= 122
o
F and injection rate 4850 lbm/hr with using black aerogel for onshore.
Reservoir pressure is clearly an important factor governing steam injectors. When
we know the formation pressure, we can determine whether we can inject steam and
at what temperatures in order to heat the reservoir. For this purpose, we could
conceive two diﬀerent formation pressures at the same depth. One is smaller than
steam pressure one is greater than steam pressure on shown Figure 6.37. If formation
pressure is greater than steam pressure, steam can not be injected to the reservoir.
Greater the steam pressures are obtained with higher temperatures. Having this
advantage, steam can be injected with increasing temperature to the formation.
84 CHAPTER 6. RESULTS AND COMPARISONS
6.5 Adding NonCondensable Gas (N
2
) in an On
shore environment
6.5.1 Examples with Insulation Materials
Figure 6.38: With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr with using black aerogel.
Noncondensable gas is added to the ﬁeld data 1 scenerio an input except 122
o
F,
time is 1 year and injection rate 4850 lbm/hr with modiﬁed Beggs and Brill twophase
correlation. First, diﬀerent N
2
molar percentage of steam is added to the steam,
starting from 5 % to 30% mol fraction of the steam. In Figure 6.38, the greater N
2
6.5. ADDINGNONCONDENSABLE GAS (N
2
) INANONSHORE ENVIRONMENT85
in the system gives more partial pressure to the system. The contribution of partial
pressure by steam drops signiﬁcantly and temperature as well. But, steam quality
drops less so the heat loss decreases. It has a disadvantage to deliver lower steam
temperature to the reservoir in terms of less latent heat addition to the reservoir.
Reducing the fraction of N
2
in the steam gives greater heat loss and quality drop,
however, it increases steam partial pressure and temperature, so more heat is delivered
to the reservoir.
Figure 6.39: With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel.
In Figure 6.39, diﬀerent injection rate is studied with 10% by mole N
2
added to
the steam. In our calculation, frictional pressure drops have the greatest contribution
86 CHAPTER 6. RESULTS AND COMPARISONS
to the pressure drop calculation. Frictional pressure drop can be caused by greater
injection rate in our case. The greater the injection rate the greater the pressure
drop, steam temperature, and steam quality decrease downhole. Increasing injection
rate does not yield good results in downward steam injection operations.
Figure 6.40: With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr with using black aerogel.
Figure 6.40, diﬀerent steam quality values are conducted in our work with 10% N
2
.
Increasing steam quality yields more pressure and temperature drop in existing cases
for noncondensable gas. However, it also yields higher steam quality values when
the steam reaches to reservoir with rich quality but poor temperature and pressure
values. That may not be good for maintaining the temperature and pressure of the
6.5. ADDINGNONCONDENSABLE GAS (N
2
) INANONSHORE ENVIRONMENT87
steam. In the same ﬁgure, heat loss values with diﬀerent steam quality does not vary
and have almost same value with diﬀerent injection quality.
Figure 6.41: With changing injection temperature, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr with using black aerogel.
Steam temperature values are changed in terms of checking the sensitivity of steam
quality, steam temperature, steam pressure, and heat loss. In Figure 6.41, increasing
steam temperature is caused by increasing steam pressure. Because of this, smaller
pressure drop and steam quality change is obtained with highest steam temperature
values. The greatest heat loss is observed with the greatest steam injection temper
ature at 650
o
F.
88 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.42: With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel.
The eﬀect of injection depth is analysed by changing the depth. However, after
2500 ft even if steam quality was high, injection depth could not be increased, because
of the signiﬁcant pressure drop along the tubing. For the diﬀerent range of depth, all
the values are corresponded and follow the same distribution except depth, shown in
Figure 6.42.
6.5. ADDINGNONCONDENSABLE GAS (N
2
) INANONSHORE ENVIRONMENT89
6.5.2 Examples without Insulation Materials
Figure 6.43: With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation.
Same scenerio is also applied for uninsulated tubing as applied to insulated tubing
cases. Figure 6.38 and Figure 6.43 has similarities in terms of the greater the mole
fraction of N
2
we have in the system, the greater pressure drop as well as temperature
drop. Another interesting observation is made with comparing to the insulated tubing
case. The two temperature and pressure drops curves have slightly higher values than
the insulated case when the N
2
mole fraction changes 5 %10%. The other curves
have greater pressure drop compaed to the insulated case. However, steam quality
drops and heat losses values converge to the same value. Steam quality is less than
90 CHAPTER 6. RESULTS AND COMPARISONS
comparing with insulated tubing, but heat loss values almost 10 times greater than
without insulated case. Using insulated tubing gives advantage instead of bare tubing
regarding to less decrease with several parameters.
Figure 6.44: With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
In Figure 6.44, diﬀerent injection rate is studied with 10% N
2
mol fraction of
noncondensable gas as shown in Figure 6.39. In our calculation frictional pressure
drops have highest contribution to the pressure drop calculation. Frictional pressure
drop can be caused by more injection rate in our case. The greater the injection
rate greater the pressure drop, steam temperature and and steam quality drop are
observed. Increasing injection rate does not yield good results in downward steam
6.5. ADDINGNONCONDENSABLE GAS (N
2
) INANONSHORE ENVIRONMENT91
injection operations. There is a greater decrease of steam quality, steam pressure,
steam temperature and greater heat losses. Similar trends are seen with insulated
tubing except more heat loss and more quality drops are calculated without insulated
tubing.
Figure 6.45: With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr without insulation.
Figure 6.45, shows diﬀerent steam quality values are conducted in our work using
uninsulated tubing with 10% N
2
. Steam quality and steam pressure or temperature
values are inversely proportional. When steam quality increase steam pressure change
decrease more as well as steam temperature in existing cases for noncondensable gas,
and vise versa. However, it also yields higher steam quality values when the steam
92 CHAPTER 6. RESULTS AND COMPARISONS
reaches to reservoir with rich quality but poor temperature and pressure values. That
may not be good for maintaining the temperature and pressure of the steam. In the
same ﬁgure, heat loss values with diﬀerent steam quality does not vary and have
almost same values with diﬀerent injection quality. Even if we have greatest liquid
water content for the lowest steam quality, the pressure drop is the smallest. The
contribution of water to the pressure drop is dependent on the slippage eﬀect value
and of course it is going to give little bit higher value than lowest water content on
density of the mixture. However, the main contribution for frictional pressure drop
values is mixture velocity of the components. In this case, the greater the steam
quality we have in the system, the greater the mixture superﬁcial velocity. This
is the main explanation for the why pressure drop is dramatically for the greatest
steam quality value. For example, I have checked three things : the total density of
the mixture, mixture velocity and frictional pressure drop values for both cases and
saw that mixture velocity is higher @ x = 0.85, than @ x = 0.6 almost two times.
Mixture density of x = 0.6 is 1.2 times higher than x = 0.85, frictional pressure drop
for x = 0.85 is almost 2.5 times higher than x = 0.6 case. Therefore, mixture velocity
contribution in twophase ﬂow calculations have highest impact on frictional pressure
drop, so total pressure drop increase more.
6.5. ADDINGNONCONDENSABLE GAS (N
2
) INANONSHORE ENVIRONMENT93
Figure 6.46: With changing injection temperature, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation.
Steam temperature values are changed in terms of checking the sensitivity of steam
quality, steam temperature, steam pressure and heat loss. In Figure 6.46, increas
ing steam temperature is caused by increasing steam pressure. Because of this less
pressure drop and steam quality change obtained with highest steam temperature
values. The greatest heat loss is observed with the greatest steam injection tempera
ture with 650
o
F. With the insulated case, it is discerned that lowest steam quality,
steam pressure, and steam temperature change found with greatest steam injection
temperature. Without insulated tubing case, it is seen that the trend of heat loss val
ues is almost 10 times greater as compared to the insulated case. Without insulated
94 CHAPTER 6. RESULTS AND COMPARISONS
case less steam quality decreases when the temperature 550
o
F, the other higher two
values have corresponding values.
Figure 6.47: With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
Changing depth under noncondensable gas with using uninsulated tubing is stud
ied by changing the depth ranging from 1600 ft to 2500 ft. Again, a similar trend
is shown in Figure 6.47 using insulated tubing. However, even if values seems closer
for the two diﬀerent cases, heat loss values are always much greater as compared to
the insulated case.
6.6. ADDINGNONCONDENSABLE GAS (N
2
) INANOFFSHORE ENVIRONMENT95
6.6 Adding NonCondensable Gas (N
2
) in an Oﬀ
shore environment
Similar observation of steam temperature, pressure, quality and heat loss are obtained
with onshore environment, except for initial pressure drop is greater for oﬀshore
environment in both insulated and uninsulated cases. For example, Figure 6.38 and
Figure 6.48 steam pressure, steam temperature, steam quality are similar except
heat losses, on Figure 6.48 heat loss values corresponds on one line, Figure 6.39 and
Figure 6.49, Figure 6.40 and Figure 6.50, Figure 6.41 and Figure 6.51, and Figure
6.42 and Figure 6.52 have similar trends between onshore cases and oﬀshore cases
with insulated tubing. This observation continues for uninsulated tubing cases as
well. For instance, Figure 6.43 and Figure 6.53, Figure 6.44 and Figure 6.54, Figure
6.45 and Figure 6.55, Figure 6.46 and Figure 6.56, and Figure 6.47 and Figure 6.57
have similar trends between onshore cases and oﬀshore cases with uninsulated tubing.
6.6.1 Examples with Insulation Materials
96 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.48: With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr with using black aerogel.
Figure 6.49: With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel.
6.6. ADDINGNONCONDENSABLE GAS (N
2
) INANOFFSHORE ENVIRONMENT97
Figure 6.50: With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr with using black aerogel.
Figure 6.51: With changing injection temperature, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr with using black aerogel.
98 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.52: With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr with using black aerogel.
6.6. ADDINGNONCONDENSABLE GAS (N
2
) INANOFFSHORE ENVIRONMENT99
6.6.2 Examples without Insulation Materials
Figure 6.53: With changing N
2
molar percentage, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation.
Figure 6.54: With changing injection rate, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
100 CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.55: With changing steam quality molar percentage , 1 year, T
m
= 122
o
F
and injection rate 4850 lbm/hr without insulation.
Figure 6.56: With changing injection temperature, 1 year, T
m
= 122
o
F and injection
rate 4850 lbm/hr without insulation.
6.6. ADDINGNONCONDENSABLE GAS (N
2
) INANOFFSHORE ENVIRONMENT101
Figure 6.57: With changing injection depth, 1 year, T
m
= 122
o
F and injection rate
4850 lbm/hr without insulation.
Chapter 7
Summary, Conclusions and Future
Work
7.1 Summary
In this thesis, we have presented an approximate, but accurate, solution for heat losses
from steam injectors in both onshore and oﬀshore environments. Solutions take into
account pressure drop, heat loss, and multiphase ﬂow. We compared our solution
with two ﬁeld cases and got good agreement. We improved the solution provided by
Fontanilla[19] by using more accurate correlations. We made our oﬀshore scenerio
calculation taking into account seawater thermo physical properties in order to create
a more realistic solution.
We used properties that are characteristics of six diﬀerent insulation materials that
are 1) black aerogel, 2) white aerogel, 3) ﬁberglass, 4) carbon ﬁber, 5) thermolastic
insulation and 6) calcium silicate [34] with four diﬀerent correlations for frictional
pressure drop in the tubing: Aziz et al. [29], Beggs and Brill [3, 7, 14, 39], Hasan
and Kabir [23, 37, 40] , and Drift Flux[21] and gave the results in results section.
When the thermal conductivity of the insulation material increases, pressure drop
102
7.2. CONCLUSIONS 103
increases, steam quality drops, and we lose more heat to the surroundings. Heat loss
and steam quality are highly related with the steam temperature and steam pressure.
In order to see this, we used three diﬀerent wellhead temperatures of 400
o
F, 500
o
F,
and 600
o
F and plotted the results.
In our calculations, although we saw that Beggs and Brill gives good approx
imation with the ﬁeld data 12 in temperature, pressure calculations obtained over
predicted steam quality results based on Fontanilla’s approach. As we know each well
has unique characteristics and production or completions history that makes predic
tions a little bit harder. Therefore, in order to get more accurate simulation results,
more published ﬁeld data is needed.
7.2 Conclusions
1. Heat transfer equations are developed by making an analogy with circuits for
oﬀshore application and simple heat loss calculation is done with insulated and
uninsulated tubing. Uninsulated tubing gives 8 times higher heat loss when
compared to the worst insulation material and 30 times more heat loss with
compared to the best insulation material. The lower the thermal conductivity,
the lesser the heat loss to the surroundings.
2. The Fontanilla and Aziz approach is augmented and tested against the two ﬁeld
cases. With improved twophase ﬂow correlations, a better estimation of tem
perature proﬁle is obtained with the Fontanilla and Aziz approach. Pressure
and temperature distribution values converged to the ﬁeld data 1 and 2, with
using the Beggs and Brill model. However, Beggs and Brill overpredicts steam
quality values with compared to Fontanilla’s results. The other twophase corre
lations give similar values with Fontanilla’s results in terms of comparing steam
quality. Four diﬀerent twophase ﬂow correlations yield almost same heat loss
104 CHAPTER 7. SUMMARY, CONCLUSIONS AND FUTURE WORK
results in both ﬁeld cases.
3. Field data of steam quality values are not available in the literature. The
calculated steam quality values based on all the investigated twophase ﬂow
correlations are compared with the results of Fontanilla and Aziz. The Beggs
and Brill model yields higher steam quality values, while all the other models
yield similar results with the base Fontanilla and Aziz calculation.
4. Using insulation materials only aﬀected heat losses and steam quality for both
onshore and oﬀshore environments using insulated or uninsulated tubing. Hav
ing low thermal conductive material such as black aerogel provides signiﬁcantly
less heat loss and steam quality decrease over the length of the well. Onshore
and oﬀshore results for both insulated or uninsulated cases showed similar trends
in steam quality, steam pressure, steam temperature and heat loss proﬁles.
5. A GUI is developed for users to make the calculation faster, and see the other
parameters eﬀect on steam injection operations.
6. Formation pressure is important for both onshore and oﬀshore steam injection
operations. One case is studied for two diﬀerent environments, and it is shown
that when the formation pressure is greater than steam injection pressure at
given depth steam can not be delivered to the reservoir. Therefore, formation
pressure importance is emphasized in Figure 6.37.
7. The use of noncondensable gas, N
2
, is added to steam is explored. Sensitivity
analysis of the diﬀerent parameters are conducted. Identical wellhead steam
injection temperature values for oﬀshore insulated tubing are considered with
addition of 10 % N
2
to steam and a pure steam only case. We observed that
having 10 % N
2
increases steam quality value around 8 %, decreases saturated
7.3. FUTURE WORK 105
steam pressure around 20 %, and decreases heat loss around 2 %. For uninsu
lated tubing cases, increasing N
2
mole fraction does not change the bottomhole
steam quality and yielded greater overall heat loss. The trend with diﬀerent non
condensable gas mole fraction, diﬀerent steam quality, diﬀerent injection rate,
diﬀerent depth and diﬀerent injection temperature values are similar both insu
lated and uninsulated cases in for onshore and oﬀshore environments. Greater
the steam injection rate caused greater overall heat loss. Changing depth do not
have any contribution on both insulated tubing and uninsulated tubing cases.
The greatest eﬀect on bottomhole conditions of steam quality is observed with
the highest wellhead steam temperature. Increasing steam temperature leads to
greater heat loss, because the temperature diﬀerence between the well and the
formation is greater. At the same time, the greater the injection temperature
the greater the pressure values and the smaller the change on steam pressure
and steam temperature.
7.3 Future Work
For future work, in order to expand of understanding of the physics of the problem,
more ﬁeld data has to be provided by the industry. With this more accurate results
can be obtained. Sensitivity analysis and optimization can be done with diﬀerent
injection temperature, rate, time, steam quality and depth, in order to ﬁnd best
combination with regarding cost constraints and technology.
106 CHAPTER 7. SUMMARY, CONCLUSIONS AND FUTURE WORK
Nomenclature
Abbreviations
2ResWithIns Sea ﬂoor to the reservoir with insulation
2ResWithoutIns Sea ﬂoor to the reservoir without insulation
SeaWithIns Oﬀshore cases with insulation
SeaWithoutInsOﬀshore cases without insulation
SLWithIns Surface line with insulation
SLWithoutIns Surface line without insulation
HSE health, safety and environment
VIT vacuum insulated tubing
Symbols
a geothermal gradient,
o
F/ft
A area,ft
2
B(θ) inclination factor
C
L
noslip liquid holdup
˙
Q
ls
rate of heat loss per unit length of pipe in
dp pressure change, psi
dQ heat loss to the surrounding, BTU/lb
dq heat loss to the surrounding, BTU/lb
E
L
(0) horizontal liquid holdup
E
L
(θ) actual liquid holdup
f friction factor, dimensionless
f(tD) time conduction function also known as Ramey function
f
tp
twophase friction factor, dimensionless
7.3. FUTURE WORK 107
f
NS
no slip friction factor, dimensionless
F
rm
Froud number of mixture
g acceleration due to gravity, 4.17e −8 ft/hr
2
Gr Grashoﬀ’s number
h enthalpy, BTU/lb
h
f
ﬁlm coeﬃcient of heat transfer of the pipe, BTU/ft
2
−hr
h
fc
coeﬃcient of heat transfer forced convection, , BTU/ft
2
−hr
h
pipe
coeﬃcient of the heat transfer of pipe, BTU/ft
2
−hr
h
c,an
radiation and convection coeﬃcient of heat transfer, BTU/ft
2
−hr
H
L
liquid holdup density
J mechanical equivalent of heat, 778 ft −lb
f
/BTU
L length, ft
k thermal conductivity of the material, BTU/(ft −hr −
o
F)
khc eﬀective thermal conductivity of the annular ﬂuid, BTU/(ft −hr −
o
F)
kha actual thermal conductivity of the annular ﬂuid, BTU/(ft −hr −
o
F)
KE kinetic energy, BTU/lb
N
Re
Reynolds number
P pressure, psi
PE potential energy, BTU/lb
Pr Prandtl’s number
q
g
gas ﬂow rate, ft
3
/hr
r
ti
inner radius of the tubing, ft
r
to
outer radius of the tubing, ft
r
ins
insulation radius of the tubing, ft
r
ci
inside radius of the casing, ft
r
co
outside radius of the casing, ft
r
h
wellbore radius, ft
108 CHAPTER 7. SUMMARY, CONCLUSIONS AND FUTURE WORK
r
Ea
radius of the altered zone in the earth near the well, ft
R Reynolds number
R
h
speciﬁc thermal resistance
R
NS
noslip Reynolds number
t time, hrs
T temperature,
o
F
Ta absolute temperature,
o
R =
o
F + 460
T
A
ambient temperature of the atmosphere,
o
F
T
b
bulk temperature of the ﬂuid in the pipe,
o
F
Tm mean surface temperature,
o
F
U overall coeﬃcient of heat transfer, BTU/hr −ft
2
−
o
F
v speciﬁc volume, ft
3
/lb
v
w
wind velocity, mph
V velocity, ft/hr
v
sg
superﬁcial velocity for gas phase, ft/hr
v
g
actual velocity for gas phase, ft/hr
v
sL
superﬁcial velocity for liquid phase, ft/hr
v
s
actual velocity for liquid phase, ft/hr
v
m
mixture velocity, ft/hr
Wm steam injection rate, lb/hr
X steam quality, fraction by weight
z elevation or depth, ft
Greek Symbols
α
E
thermal diﬀusivity of the earth, ft
2
/hr
blackbody emissive power, BTU/hr −ft
2
7.3. FUTURE WORK 109
ins
emissivity of the insulation material , BTU/hr −ft
2
ci
emissivity of the casing , BTU/hr −ft
2
λ
pipe
thermal conductivity of the pipe,BTU/(ft −hr −
o
F)
λ
ins
thermal conductivity of the insulation,BTU/(ft −hr −
o
F)
λ
cem
thermal conductivity of the cement,BTU/(ft −hr −
o
F)
λ
Ea
thermal conductivity of the altered zone,BTU/(ft −hr −
o
F)
λ
E
thermal conductivity of the unaltered zone,BTU/(ft −hr −
o
F)
µ
s
gas viscosity,lb/ft −hr
µ
w
liquid viscosity,lb/ft −hr
ρ
l
gas density, lbm/ft
3
ρ
l
liquid density, lbm/ft
3
ρ
s
slip mixture density, lbm/ft
3
ρ
NS
noslip mixture density, lbm/ft
3
σ Stefan Boltzman constant, 0.1714e −8 BTU/(hr −ft
2
−
o
R
4
)
θ angle from horizontal, −90
Appendix A
Derivation of the Equations
A.1 Total Energy Equation
In Appendix A, all the equations are listed that are adapted and modiﬁed from the
work that was done by Fontanilla and Aziz [19]. For a constant injection rate the
continuity equation may be written as
d
dz
ρ
tp
V
m
r
2
ti
= 0 (A.1.1)
where, ρ
tp
is the twophase density (lb/ft
3
), V
m
is the mixture velocity (ft/hr), and
r
ti
is the inside tubing radius (ft).
The energy conservation equation may be derived by taking a diﬀerential element
∆z of the tubing as in Figure A.1:
Energy In (BTU/lb) = Energy Out (BTU/lb)
h
m
1
+
z
1
J
+
V
2
m
1
2gJ
= h
m
2
+
z
2
J
+
V
2
m
2
2gJ
(A.1.2)
where, h
m
1
is the mixture enthalpy at point 1 (BTU/lb), z
1
is the elevation at point
1 (ft), J is the mechanical equivalent of heat = 778 (ftlb
f
/BTU), V
m
1
is the mixture
110
A.1. TOTAL ENERGY EQUATION 111
velocity at point 1 (ft/hr), g is the acceleration due to the gravity = 4.17*10
8
(ft/hr
2
),
h
m
2
is the mixture enthalpy at point 2 (BTU/lb), z
2
is the elevation at point 2 (ft),
V
m
2
is the mixture velocity at point 2 = V
m
1
+dV
m
(ft/hr), dQ is the heat loss to the
surroundings, BTU/lb.
Figure A.1: Schematic view of tubing element in our calculation.
Transferring the L.H.S. of Eq. A.1.2 to the R.H.S. and assuming dV
2
m
is negligible
we have,
h
m
2
−h
m
1
+
z
2
−z
1
J
+
V
m
1
dVm
gJ
+dQ = 0 (A.1.3)
∆Enthalpy + ∆PE + ∆KE + Heat Loss to Surrounding = 0
The sign of the second term, ∆PE, is correct only if the elevation decreases downward.
In the present problem for the seaﬂoor to reservoir, the surface is the datum plane
and the elevation increases downward, so we need to aﬃx a negative sign before the
112 APPENDIX A. DERIVATION OF THE EQUATIONS
term.
dh
m
+
dz
J
+
V
m
1
dVm
gJ
+dQ = 0 (A.1.4)
dividing by dz
dh
m
dz
+
1
J
+
V
m
1
gJ
+
dV
m
dz
+
dQ
dz
= 0 (A.1.5)
The kinetic energy term in Eq. A.1.5 may be expanded as :
V
m
1
= V
SL
+V
SG
= G
L
v
L
+ G
g
v
g
(A.1.6)
where, V
SL
is the superﬁcial liquid velocity (ft/hr), V
SG
is the superﬁcial gas velocity
(ft/hr), G
L
is the liquid mass ﬂux rate (lb/hrft
2
), v
L
is the speciﬁc volume of liquid
(ft
3
/lb), G
g
is the gas mass ﬂux rate (lb/hrft
2
), v
g
is the speciﬁc volume of gas
(ft
3
/lb). Its equivalent diﬀerential form is
dV
m
= G
L
dv
L
+G
g
dv
g
(A.1.7)
Therefore we have
V
m
1
gJ
dV
m
dz
=
1
gJ
_
v
L
G
2
L
dV
L
dz
+v
L
G
L
G
g
dV
g
dz
+ v
g
G
g
G
L
dV
L
dz
+v
g
G
2
g
dV
g
dz
_
(A.1.8)
Substituting Eq. A.1.8 in to Eq. A.1.5
dh
m
dz
−
1
J
+
1
gJ
+
V
m
1
gJ
+
_
v
L
G
2
L
dV
L
dz
+v
L
G
L
G
g
dV
g
dz
+v
g
G
g
G
L
dV
L
dz
+ v
g
G
2
g
dV
g
dz
_
+
dQ
dz
= 0
(A.1.9)
A.2. MECHANICAL ENERGYBALANCE OR THE EXTENDEDBERNOULLI EQUATION113
A.2 Mechanical energy balance or the Extended
Bernoulli Equation
The basis for any ﬂuid ﬂow calculation is the mechanical energy balance for the
ﬂowing ﬂuid between two points. Because there is no external work done on or by
the ﬂuid, a steadystate mechanical energy balance equation in diﬀerential form may
be written for 1 lb
m
of ﬂuid as:
144
dp
ρ
tp
+dz +
v
m
1
dv
m
g
+dW
f
= 0 (A.2.1)
where dp= P
2
P
1
is the change in pressure (lb/in
2
), ρ
tp
is the twophase density
(lb/ft
3
), dz=z
2
z
1
is the change in elevation (ft), v
m
1
is the mixture velocity at point
1 (ft/hr), dv
m
is the change in mixture velocity (ft/hr), g is the acceleration due
to gravity = 4.17*10
8
(ft/hr
2
), dW
f
is the frictional drag (ft). We again have to
aﬃx a negative sign on the second term because the elevation increases downward.
Multiplying Eq. A.2.1 by
ρtp
dz
, we have
144
dp
dz
−ρ
tp
+
ρ
tp
v
m
1
g
dv
m
dz
+ρ
tp
dW
f
dz
= 0 (A.2.2)
or 144
dp
dz

dp
dz
elevation of PE+
dp
dz
acceleration or KE+
dp
dz
friction = 0
Substituting Eq. A.1.9 into Eq. A.2.2, we have
144
dp
dz
−ρ
tp
+
ρ
tp
g
_
v
L
G
2
L
dV
L
dz
+v
L
G
L
G
g
dV
g
dz
+v
g
G
g
G
L
dV
L
dz
+ v
g
G
2
g
dV
g
dz
_
+
_
dP
dz
_
friction
(A.2.3)
The mixture enthalpy h
m
is a function of steam quality X and pressure P (psia).
h
m
= f(X, P) (A.2.4)
114 APPENDIX A. DERIVATION OF THE EQUATIONS
The speciﬁc volume of liquid (water) and gas (steam) are just functions of pressure
P (psia). V
L
= g (P),V
g
= h (P) thus we can write
dh
m
dz
=
∂h
m
∂X
dX
dz
+
∂h
m
∂P
dP
dz
(A.2.5)
dv
g
dz
=
∂v
g
∂P
dP
dz
(A.2.6)
dv
L
dz
=
∂v
L
∂P
dP
dz
(A.2.7)
Substituting Eq. A.2.5, Eq. A.2.6, and Eq. A.2.7 into Eq. A.1.9 yields:
∂h
m
∂X
dX
dz
−
∂h
m
∂P
dP
dz
−
1
J
+
1
gJ
dP
dz
_
v
L
G
2
L
∂V
L
∂P
+ v
L
G
L
G
g
∂V
g
∂P
+ v
g
G
g
G
L
∂V
L
∂P
+ v
g
G
2
g
∂V
g
∂z
_
+
dQ
dz
= 0
(A.2.8)
Substituting Eq. A.2.6, and Eq. A.2.7 into Eq. A.2.3 we have:
144
dP
dz
−ρ
tp
+
ρ
tp
g
dP
dz
_
v
L
G
2
L
∂V
L
∂P
+v
L
G
L
G
g
∂V
g
∂P
+v
g
G
g
G
L
∂V
L
∂P
+ v
g
G
2
g
∂V
g
∂P
_
+
_
dP
dz
_
friction
= 0
(A.2.9)
Solving for
dP
dz
in Eq. A.2.9,
dP
dz
=
ρ
tp
−
_
dP
dz
_
friction
144 +
ρtp
g
_
v
L
G
2
L
∂V
L
∂P
+v
L
G
L
G
g
∂Vg
∂P
+v
g
G
g
G
L
∂V
L
∂P
+v
g
G
2
g
∂Vg
∂P
_ (A.2.10)
Solving for
dX
dz
in Eq. A.2.8 yields,
dX
dz
=
1
dhm
dX
_
−
∂h
m
∂P
dP
dz
+
1
J
−
1
gJ
dP
dz
_
v
L
G
2
L
∂V
L
∂P
+v
L
G
L
G
g
∂V
g
∂P
+v
g
G
g
G
L
∂V
L
∂P
+v
g
G
2
g
∂V
g
∂P
_
−
dQ
dz
_
(A.2.11)
Eq. A.2.10 and Eq. A.2.11 are two simultaneous, ﬁrst order ordinary diﬀerential
equations to be solved for pressure (P) and quality (X). Besides these we need to ﬁnd
A.3. EVALUATION OF HEAT LOSS TO THE SURROUNDING 115
a method to calculate
dQ
dz
.
A.3 Evaluation of Heat Loss to the Surrounding
In order to solve Eq. A.2.11, we need to know
dQ
dz
(BTU/lbft) which is the heat loss
to the surroundings.
dQ =
dq
W
m
(A.3.1)
dQ is the change in heat loss by ﬂuid (BTU/lb), dq is the change in heat loss rate
(BTU/hr), W
m
is the steam injection rate (lb/hr)
Because we have a pseudo steady state heat ﬂow conditions in the wellbore, the
rate of heat conduction from the ﬂuid to cementformation interface (hole) is ex
pressed:
dq
dz
= 2πr
to
(T
f
−T
h
) (A.3.2)
A representation of the problem for three parts of our system will be analyzed and we
refer those ﬁgures here again. Figure 3.1 refers to surface lines, Figure 3.4 refers to
sea level to sea ﬂoor and Figure 3.9 refers to sea ﬂoor to reservoir heat loss calculation
with insulated or not. Three ﬁgures and having two options each of them gives six
diﬀerent heat loss calculation procedures. We adapted the surface line heat losses
from Prats and the rest of theprocedures are based on Willhite’s approach that was
discussed in chapter 3. The radius r and overall heat transfer coeﬃcient U used in Eq.
A.3.2 is based on any reference point. In the present case, the outer tubing surface
was chosen as the reference plane. From Ramey[42], the rate of heat conduction into
the earth is expressed as,
dq
z
=
2πk
e
(T
h
−T
e
)
f(t)
(A.3.3)
where: T
e
is the temperature of the earth (
o
F), k
e
is the thermal conductivity of the
116 APPENDIX A. DERIVATION OF THE EQUATIONS
earth (BTU/hrft
o
F), f(t) is the time conduction function.
Because the rate of heat conduction from the ﬂuid to the cement formation inter
face (hole) must equal the rate of heat conduction into the earth, Eq. A.3.2 and Eq.
A.3.3 may be equated to obtain an expression for T
h
:
T
h
=
r
to
U
to
f(t)T
f
+ k
e
T
e
r
to
U
to
f(t) + k
e
(A.3.4)
Substituting Eq. A.3.4 and Eq. A.3.1 into Eq. A.3.2 we have:
dQ
dz
=
2πr
to
U
to
W
m
_
T
f
r
to
U
to
f(t)T
f
+k
e
T
e
r
to
U
to
f(t) + k
e
_
(A.3.5)
If the geothermal gradient is provided, the temperature of the earth, T
e
, is given by
T
e
=T
m
+az, otherwise it assumes constant value. T
m
is the mean surface temperature
(
o
F), a is the geothermal gradient ((
o
F)/ft), z is the depth (ft). When we substitute
the temperature of the earth into Eq. A.3.5 we have:
dQ
dz
=
2πr
to
U
to
W
m
_
T
f
r
to
U
to
f(t)T
f
+k
e
(T
m
+az)
r
to
U
to
f(t) + k
e
_
(A.3.6)
A.4 Determination of the U
to
and T
ci
To obtain
dQ
dz
(Eq. A.3.6), we need to know U
to
which is the overall heat transfer
coeﬃcient or the overall coeﬃcient of conductance. The overall heat transfer con
ductance U
to
may be evaluated by writing expressions for the heat ﬂow through the
various components of the injector. The rate of heat conductance from the ﬂuid to
the cementformation (hole) interface is given by A.3.2. The rate of the heat transfer
between the ﬂowing ﬂuid and the inside tubing is given as:
dq
dz
= 2πr
ti
h
f
(T
f
−T
i
) (A.4.1)
A.4. DETERMINATION OF THE U
TO
AND T
CI
117
where h
f
is the ﬁlm coeﬃcient for heat transfer based on the inside surface area of the
tubing. Heat ﬂow through the tubing wall, insulation, casing wall, and the cement
sheath occurs by conduction. Integrating Fourier’s law of heat conduction over the
thickness of the tubing wall we write:
dq
dz
=
2πk
tub
ln(
rto
r
ti
)
(T
ti
−T
to
) (A.4.2)
Integrating Fourier’s law of heat conduction over the thickness of the insulation yields:
dq
dz
=
2πk
ins
ln(
r
ins
rto
)
(T
to
−T
ins
) (A.4.3)
There are three modes of heat transfer present in the casing annulus.
1. heat conduction through the ﬂuid in the annular space.
2. natural convection of ﬂuid
3. radiation
Heat transfer by natural convection in the annulus is caused by ﬂuid motion result
ing from the variation of density with temperature. Hot ﬂuid near the tubing and
insulation is less dense than the ﬂuid in the center of the annulus and tends to rise.
On the other hand, the ﬂuid near the casing wall is cooler and denser than that in
the center of the annulus and tends to fall. When a body is heated, radiant energy is
emitted at a rate dependent on the temperature of the body. The amount of radiant
energy transported between the tubing/insulation and casing depends on the view
the surfaces have of each other and their emitting and absorbing characteristics. To
account for these three modes of heat transfer, it is convenient to deﬁne the heat rate
through the annulus in terms of heat transfer coeﬃcients h
c
(natural convection and
118 APPENDIX A. DERIVATION OF THE EQUATIONS
conduction) and h
r
(radiation), as:
dq
dz
= 2πr
to
(h
c
+ h
r
)(T
ins
−T
ci
) (A.4.4)
The heat conduction through the casing wall is given as:
dq
dz
=
2πk
cas
ln(
rco
r
ci
)
(T
ci
−T
co
) (A.4.5)
The heat conduction throught the cement sheath is given as:
dq
dz
=
2πk
cem
ln(
r
h
rco
)
(T
co
−T
h
) (A.4.6)
Note that:
T
f
−T
h
= (T
f
−T
ti
)+(T
ti
−T
to
)+(T
to
−T
ins
)+(T
ins
−T
ci
)+(T
ci
−T
co
)+(T
co
−T
h
) (A.4.7)
Solving for the temperature diﬀerences
T
f
−T
h
=
dq
dz
2πr
to
U
to
(A.4.8)
T
f
−T
ti
=
dq
dz
2πr
ti
h
f
(A.4.9)
T
ti
−T
to
=
dq
dz
2πk
tub
ln
_
r
to
r
ti
_
(A.4.10)
T
to
−T
ins
=
dq
dz
2πk
ins
ln
_
r
ins
r
to
_
(A.4.11)
T
ins
−T
ci
=
dq
dz
2πr
to
(h
c
+h
r
)
(A.4.12)
A.4. DETERMINATION OF THE U
TO
AND T
CI
119
T
ci
−T
co
=
dq
dz
2πk
cas
ln
_
r
co
r
ci
_
(A.4.13)
T
co
−T
h
=
dq
dz
2πk
cem
ln
_
r
h
r
co
_
(A.4.14)
Substituting Eqs. A.4.9  4.14 into the Eq. A.4.7 we will have:
T
f
−T
h
=
dq
dz
2π
_
_
1
r
ti
h
f
+
ln
_
rto
r
ti
_
k
tub
+
ln
_
r
ins
rto
_
k
ins
+
1
r
to
(h
c
+h
r
)
+
ln
_
rco
r
ci
_
k
cas
+
ln
_
r
h
rco
_
k
cem
_
_
(A.4.15)
An equation for U
to
is found by comparing Eq. A.4.8 and Eq. A.4.15
1
r
to
U
to
=
_
_
1
r
ti
h
f
+
ln
_
rto
r
ti
_
k
tub
+
ln
_
r
ins
rto
_
k
ins
+
1
r
to
(h
c
+h
r
)
+
ln
_
rco
r
ci
_
k
cas
+
ln
_
r
h
rco
_
k
cem
_
_
(A.4.16)
Therefore,
U
to
=
_
_
r
to
r
ti
h
f
+
r
to
ln
_
rto
r
ti
_
k
tub
+
r
to
ln
_
r
ins
rto
_
k
ins
+
1
(h
c
+ h
r
)
+
r
to
ln
_
rco
r
ci
_
k
cas
+
r
to
ln
_
r
h
rco
_
k
cem
_
_
−1
(A.4.17)
The thermal conductivity of the tubing and casing steel (k
tub
= k
cas
= 25BTU/hr −
ft −
o
F) is considerably higher than that for the other materials in the wellbore
(insulation = 0.02 to 0.06 BTU/hrft
o
F and cement = 0.2 to 0.6 BTU/hrft
o
F).
Therefore, its relative contribution in Eq. A.4.17 is small and can be neglected (T
ti
=
T
to
and T
ci
= T
co
).
120 APPENDIX A. DERIVATION OF THE EQUATIONS
Also the ﬁlm coeﬃcient h
f
for steam and water are high enough (500 to 4000
BTU/hrft
2

o
F) to justify the assumption of inﬁnite ﬁlm coeﬃcient (T
f
= T
ti
). Thus
Eq. A.4.17 simpliﬁes to Eq. A.4.18
U
to
=
_
_
r
to
ln
_
r
ins
rto
_
k
ins
+
1
(h
c
+h
r
)
+
r
to
ln
_
r
h
rco
_
k
cem
_
_
−1
(A.4.18)
A.5 Determination of the Convection Heat Trans
fer Coeﬃcient
Before U
to
can be calculated in Eq. A.4.18, the convection coeﬃcient h
c
and the
radiation h
r
must be evaluated. Heat transfer per unit length by conduction and free
convection in the annulus is given if there is no insulation;
dq
c
dz
=
2πk
hc
ln
_
r
ci
rto
_ (A.5.1)
if there is an insulation;
dq
c
dz
=
2πk
hc
(T
ins
−T
ci
)
ln
_
r
ci
rto
_ (A.5.2)
We can also express Eq. A.5.1 and Eq. A.5.2 may also be expressed if there is no
insulation as:
dq
c
dz
= 2πr
to
h
c
(T
to
−T
ci
) (A.5.3)
if there is a insulation:
dq
c
dz
= 2πr
ins
h
c
(T
ins
−T
ci
) (A.5.4)
A.5. DETERMINATIONOF THE CONVECTIONHEAT TRANSFER COEFFICIENT121
Combining Eq. A.5.1 with Eq. A.5.3 and Eq. A.5.2 with Eq. A.5.4 we will have if
there is no insulation,
h
c
=
khc
r
to
ln
_
r
ci
rto
_ (A.5.5)
if insulation present;
h
c
=
khc
r
ins
ln
_
r
ci
r
ins
_ (A.5.6)
The eﬀective thermal conductivity of the annular ﬂuid (khc) is related to the actual
thermal conductivity of the annular ﬂuid (kha) as a function of the Grashoﬀ number
and Prandtl number.
khc = (kha)(0.049)(GrPr)
0.333
(Pr)
0.074
(A.5.7)
where: Gr is the Grashoﬀ’s number, if there is no insulation
Gr =
(r
ci
−r
to
)
3
gρ
2
an
β(T
to
−T
ci
)
µ
2
an
(A.5.8)
if there is an insulation
Gr =
(r
ci
−r
ins
)
3
gρ
2
an
β(T
ins
−T
ci
)
µ
2
an
(A.5.9)
Pr is the Prandtl’s number
Pr =
C
an
µ
an
kha
(A.5.10)
where C
an
is the speciﬁc heat of the annular ﬂuid and µ
an
is the viscosity of the
annular ﬂuid with given pressure in the annulus. β (
o
R
−1
) is the coeﬃcient of volume
expansion. For ideal gas, β is the reciprocal average absolute annulus temperature.
β =
1
(T
an
+ 460)
(A.5.11)
122 APPENDIX A. DERIVATION OF THE EQUATIONS
where: T
an
if there is no insulation
T
an
=
T
to
+T
ci
2
(A.5.12)
if there is an insulation
T
an
=
T
ins
+T
ci
2
(A.5.13)
A.6 Determination of the Radiation Heat Transfer
Coeﬃcient
Radiation heat transfer problems can be represented by ”radiation networks” as in
Carslaw and Jaeger and Herrera, et al.. The following developments are only for
bare tubings but an equivalent equation which can be used for calculation of h
r
for tubings with insulation is also given. The tubing and the casing surfaces which
exchange heat with each other may be modelled by resistances in series as shown in
Fontanilla’s thesis [19]. Here we are going to give the equations for h
r
with insulation
and without insulation. So h
r
without insulation
h
r
=
(Ta
to
+Ta
ci
)(Ta
2
to
+Ta
2
ci
)σ
1
to
+
rto
r
ci
_
1
ci
−1
_ (A.6.1)
h
r
with insulation
h
r
=
(Ta
ins
+Ta
ci
)(Ta
2
ins
+Ta
2
ci
)σ
1
ins
+
r
ins
r
ci
_
1
ci
−1
_ (A.6.2)
A.7 Computational Procedure for U
to
In order to calculate h
c
, h
r
and hence U
to
, we need to know the temperature of the
inside casing T
ci
and the temperature at the outer insulation T
ins
. Since we need to
A.7. COMPUTATIONAL PROCEDURE FOR U
TO
123
know T
ci
and T
ins
(which are unknown) to solve for h
c
, h
r
and U
to
, a trial and error or
an iterative solution is required to determine the proper combination of U
to
, T
ci
and
T
ins
. Before describing the iteration procedure, the equations to be used the iteration
are developed. The equation for T
ins
is obtained from Eq. (A.4.3):
T
ins
= T
to
−
dq
dz
ln
_
r
ins
rto
_
2πk
ins
(A.7.1)
The equation for T
co
is obtained from Eq. (A.4.6):
T
ins
= T
h
+
dq
dz
ln
_
r
h
rco
_
2πk
cem
(A.7.2)
Substituting this expression for T
co
in Eq. (A.4.5), and solving for T
ci
, we have:
T
ci
= T
h
+
dq
dz
ln
_
rco
r
ci
_
2πk
cas
+
dq
dz
ln
_
r
h
rco
_
2πk
cem
(A.7.3)
Substituting Eq. (A.3.2) for
dq
dz
in Eq. (A.7.3) and solving for T
ci
T
ci
= T
h
+r
to
U
to
(T
f
−T
h
) +
_
ln
rco
r
ci
k
cas
+
ln
r
h
rco
k
cem
_
(A.7.4)
By negleccting the thermal resistance of the ﬁlm, tubing, and casing, Eq. (A.7.4)
reduces to :
T
ci
= T
h
+
r
to
U
to
ln
r
h
rco
k
cem
(T
to
−T
h
) (A.7.5)
The hole temperature (T
h
) can be obtained from Eq. (A.3.4) derived earlier.
T
h
=
r
to
U
to
f(t)T
f
+ k
e
T
e
r
to
U
to
f(t) + k
e
(A.7.6)
124 APPENDIX A. DERIVATION OF THE EQUATIONS
The ﬁrst thing to do is to assume an arbitrary value for
dq
dz
and calculate an initial T
ins
from Eq. (A.7.1). Then set T
ci
equal to the geothermal temperature (T
e
). Having
values for T
ins
and T
ci
; h
r
, h
c
, U
to
, and f(t) can be calculated. T
h
is then solved using
Eq. (A.7.6). Then a new value of T
ci
is calcualted using Eq. (A.7.5). The old and
new casing temperatures are compared at this point. If the absolute value of their
diﬀerence is greater than a tolerable amount, say 1
o
F, the old casing temperature
is incremented by 70 percent of the diﬀerence. The model then uses this new casing
temperature. To obtain a corresponding T
h
based on the new T
ci
, solving for T
h
in
Eq. (A.7.6) yields:
T
h
=
T
ci
k
cem
−r
to
ln
r
h
rco
T
to
k
cem
−r
to
U
to
ln
r
h
rco
(A.7.7)
Using the new T
h
from Eq. (A.7.7), we calculate
dq
dz
from Eq. (A.3.2). Calculate the
new T
ins
from Eq. (A.7.1) using the new
dq
dz
, T
ins
and T
ci
. The iteration is continued
until convergence is obtained.
In summary, the iteration procedure is as follows:
1. Assume an arbitrary value for old
dq
dz
and calculate the old T
ins
from Eq. (A.7.1).
2. Set T
ci
equal to the geothermal temperature T
e
and call this the old T
ci
3. Calculate h
r
from Eq. (A.6.2), h
c
from Eq. (A.5.6) U
to
from Eq. (A.4.18) and
f(t) from Eq. (A.8.1) knowing the old T
ins
and the old T
ci
4. Calculate the old T
h
from Eq. (A.3.4)
5. Calculate the new T
ci
using Eq. (A.7.5).
6. The old T
ci
and new T
ci
are compared. if newT
ci
 oldT
ci
 > 1, the old T
ci
is
incremented by 70 percent of the absolute value of the diﬀerence. Otherwise
convergence is obtained.
A.8. DETERMINATION OF F(T) 125
7. If convergence is not yet obtained, a corresponding T
h
based on the T
ci
incre
mented in step 6. is calculated using Eq. (A.7.7).
8. Calculate a corresponding
dq
dz
from Eq. (A.3.2).
9. Calculate T
ins
from Eq. (A.7.1)
10. Starting with step 3. the procedure is repeated until convergence is obtained.
A.8 Determination of f(t)
The time conduction function f(t) introduced in the equation of unsteady state heat
ﬂow to the earth and needed to obtain
dQ
dz
and U
to
can be estimated from solutions for
radial heat conduction from an inﬁnitely long cylinder. Such solutions are analogous
to transient ﬂuid ﬂow solutions used in reservoir engineering. As can be seen Figure
A.5 all ﬁve solutions converge to the same line. the convergence time is of the order of
one week for many reservoir problems. Thus, the line source solution will often provide
useful results if times are greater than a week. From Ramey[42] an approximate
equation for f (t) satisfying the line source solution for long times is:
f(t) = −ln
_
r
h
2
√
αt
_
(A.8.1)
126 APPENDIX A. DERIVATION OF THE EQUATIONS
Figure A.2: Time conduction function (retrieved from[19]).
In this model, regression was used to obtain a third order polynomial approxima
tion for each of the ﬁve curves in Figure A.5.
Lets say K =
r
h
U
ke
the calculation procedure is as follows:
A.9. EVALUATION OF THE DERRIVATIVES 127
Algorithm 1 f(t) calculation based on the regression analysis
Require: x = log
10
_
αt
r
2
h
_
and Y = log
10
f(t)
Ensure: f(t) = 10
Y
1: if K == 0 then
2: Y ←.19865 + .48034x −.08619x
2
+.00148x
3
3: else
4: if K ≤ .2 then
5: Y ←−.12557 + .38757x −.07525x
2
+ .01065x
3
6: end if
7: if K ≤ 1.0 then
8: Y ←−.08738 + .3689x −.04619x
2
−.00222x
3
9: end if
10: if K ≤ 5.0 then
11: Y ←−.03018 + .36166x −.06586x
2
−.00393x
3
12: end if
13: if K ≥ 5.0 then
14: Y ←−.02435 + .33116x −.033723x
2
−.00525x
3
15: end if
16: end if
A.9 Evaluation of the Derrivatives
Farouq Ali[19] presented approximate correlations for speciﬁc volumes of saturated
liquid and saturated vapor.
v
L
= 0.01587 + 0.000086P
0.225
+ 0.0002P
0.45
(A.9.1)
v
L
= 363.9P
−0.9588
(A.9.2)
Diﬀerentiating with respect to pressure we have,
∂v
L
∂P
= 0.000019P
−0.775
+ 0.0009P
−0.55
(A.9.3)
∂v
g
∂P
= −348.907P
−1.9588
(A.9.4)
128 APPENDIX A. DERIVATION OF THE EQUATIONS
Farouq Ali[19] also proposed correlations for the enthalpy of saturated liquid and the
heat of vaporization L
v
h
L
= 91P
.2574
(A.9.5)
L
v
= h
g
−h
l
= 1318P
−.08774
(A.9.6)
for wet steam,
h
m
= h
L
+XL
v
(A.9.7)
h
m
= 91P
.2574
+ X(1318P
−.08774
) (A.9.8)
Diﬀerentiating with respect to the steam quality X and pressure P, we have:
∂h
m
∂X
= 1318P
−0.08774
(A.9.9)
∂h
m
∂P
= 23.423P
−0.7426
−115.64XP
−1.08774
(A.9.10)
A.10 Calculation of the Annulus Fluid Properties
There are several modiﬁcation has been done for Fontanilla[19] solutions, one of them
is getting annulus ﬂuid properties automatically. We get viscosity and thermal con
ductivity of the ﬂuid under 1 atm using ﬁgures from Prats[41].
A.10. CALCULATION OF THE ANNULUS FLUID PROPERTIES 129
Viscosity of the annular ﬂuid, µ
annulus
Figure A.3: Viscosity of the annular ﬂuid with respect to Temperature.
We took two points on our curve and showed on the ﬁgure both N
2
and air and
assumed that the line is linear with increasing temperature values, because lines are
only slightly diﬀerent from linearity. As we know from very basic way to get slope of
the line and put that value into y = mx + n equation to get the correlation between
temperature. Temperature values both N
2
and air are T
1
= 50 and T
2
= 100, and
viscosity values are µ
1
N
2
= 0.0170 cp and µ
2
N
2
= 0.0184 cp. For air same procedure
applies µ
1
air
=0.0176 cp and µ
2
air
=0.0192 cp. Let’s ﬁnd m
N
2
and m
air
values as follows:
m
N
2
=
_
µ
1
−µ
2
T
1
−T
2
_
=
_
0.0184 −0.0170
100 −50
_
= 2.8 ∗ 10
−5
(A.10.1)
130 APPENDIX A. DERIVATION OF THE EQUATIONS
m
air
=
_
µ
1
−µ
2
T
1
−T
2
_
=
_
0.0192 −0.0176
100 −50
_
= 3.2 ∗ 10
−5
(A.10.2)
Thus, slope m
N
2
= 2.8∗10
−5
and the correlation we got µ
N
2
=2.8∗10
−5
T+0.0170 and
for air µ
air
=3.2 ∗ 10
−5
T+0.0176. We used this correaltions wrt temperature in our
calculation.
Thermal conductivity of the annular ﬂuid, λ
annulus
Figure A.4: Thermal conductivity of the annular ﬂuid with respect to Temperature.
The same procedure is also applied for getting correlation both N
2
and air for
thermal conductivity with respect to temperature. Temperature values both N
2
and
air are T
1
= 400 and T
2
= 800, and thermal conductivity values are λ
1
N
2
= 0.5
BTU/ft
o
FD and λ
2
N
2
= 0.68 BTU/ft
o
FD. For air same procedure applies λ
1
air
=0.5
A.10. CALCULATION OF THE ANNULUS FLUID PROPERTIES 131
BTU/ft
o
FD and λ
2
air
=0.71 BTU/ft
o
FD. Let’s ﬁnd m
N
2
and m
air
values as follows:
m
N
2
=
_
λ
1
−λ
2
T
1
−T
2
_
=
_
0.68 −0.50
800 −400
_
= 4.5 ∗ 10
−4
(A.10.3)
m
air
=
_
λ
1
−λ
2
T
1
−T
2
_
=
_
0.71 −0.5
800 −400
_
= 5.25 ∗ 10
−4
(A.10.4)
Thus, slope m
N
2
= 4.5 ∗ 10
−4
and the correlation we got λ
N
2
=4.5 ∗ 10
−4
T+0.5 and
for air λ
air
=5.25 ∗ 10
−4
T+0.5. We used this correaltions wrt temperature in our
calculation.
Appendix B
Codes for Heat Loss Calculations
B.1 Heat Losses from Surface Line
(a) Surface pipe with insulation. (b) Surface pipe without insulation.
Figure B.1: Schematic representation of resistance to heat transfer with/without
temperature proﬁle.
132
B.1. HEAT LOSSES FROM SURFACE LINE 133
The following are the MatLab codes for our heat loss calculation starting from
surface lines with insulation and without insulation
%% Example 10. 1 Cal c ul at i ons of Heat Los s es from Sur f ace Li nes
% 4 i nc he s N−80 pi pe at r at e of 229 B/D.
% We are l ooki ng f o r the s ol ut i on at steady s t at e c ondi t i ons
% f or pi pe i s i ns ul at e d
% Thermal Recovery by Mi cheal Prats chapter 10 , pgs .125−136
%%
cl c ; cl ear al l ; cl ose al l ;
TsteamWins = 550; % F
TaverageWins = 60; % F
PipeLengthWins = 100; % f t
InjTimeWins = 365∗24; % hours
AvrWindSpeedWins = 20; % mph
r i ns i deWi ns = 0. 1478; % f t from Tabl e B. 15
routsi deWi ns = 0. 1667; % f t
r i ns ul at edWi ns = 0. 4167; % f t
lambdaPipeWins = 600/24; % Btu/ f t −hr−F
lambdaInsWins = [ 0. 166 0. 194 0. 388 0. 499 0. 569 0 . 9 6 ] . / 2 4 ;
hfWins = 2000; % Btu/sq f t −hr−F
hpiWins = i nf ; % Btu/sq f t −hr−F
hpoWins = 2000; % Btu/sq f t −hr−F
Col l ect Dat asWi t hi ns =[ ] ;
% hf c c a l c ul a t i o n based on the Eq . 10. 4
hfcWins = ( 18∗( AvrWindSpeedWins ˆ0. 6) ∗( r i ns ul at edWi ns ˆ . . .
0. 6) / r i ns ul at edWi ns ) /24;
c o l o r s = l i n e s ( length( lambdaInsWins ) ) ;
for i =1: length( lambdaInsWins )
% Over al l s p e c i f i c thermal r e s i s t a nc e
134 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
RhWins = 1/(2∗ pi ) ∗( 1/( hfWins ∗ r i ns i deWi ns ) . . .
+ 1/( hpiWins ∗ r i ns i deWi ns ) + (1/ lambdaPipeWins ) ∗ . . .
log ( routsi deWi ns / r i ns i deWi ns ) + 1/( hpoWins∗ routsi deWi ns ) +. . .
(1/ lambdaInsWins ( i ) ) ∗ log ( r i ns ul at edWi ns / routsi deWi ns ) +. . .
1/( hfcWins ∗ r i ns ul at edWi ns ) ) ;
% Heat Los s es
QlsWins = ( TsteamWins − TaverageWins )/RhWins ;
% Amount of Heat Lost from the pi pe over a per i od of ti me
QlWins = QlsWins∗InjTimeWins ∗PipeLengthWins ;
Col l ect Dat asWi t hi ns =[ Col l ect Dat asWi t hi ns ; [ RhWins QlsWins QlWins ] ] ;
Inj Ti mepl otWi ns =(1: InjTimeWins ) ’ ;
PlotQlWins =(QlsWins∗PipeLengthWins ) . ∗ Inj Ti mepl otWi ns ;
fi gure ( 1)
plot ( PlotQlWins , Inj Ti mepl otWi ns . /24 , ’ c ol or ’ , c o l o r s ( i , : ) , . . .
’ Li newi dth ’ , 3)
hold on
grid on
set ( gca , ’ XAxi sLocati on ’ , ’ top ’ , ’ YDir ’ , ’ rev ’ )
xlabel ( [ ’ Cumul ati ve Heat Los s es ( i n BTU) f or ’ . . .
num2str( PipeLengthWins ) ’ f t ’ ] , ’ Fontname ’ , ’ Cal i br i ’ , . . .
’ FontSi ze ’ , 16)
ylabel ( [ ’ Time ( i n Days ) ’ num2str( InjTimeWins / 2 4 ) ] . . .
, ’ Fontname ’ , ’ Ca l i br i ’ , ’ FontSi ze ’ , 16)
LambdaValStr = spri ntf ( ’ %0.3G’ , lambdaInsWins ( i ) ) ;
l a be l s { i } = [ ’ \lambda = ’ num2str( LambdaValStr ) ] ;
legend( l abe l s , 2 , ’ Locati on ’ , ’NE’ )
set ( gcf , ’ Uni ts ’ , ’ normal i zed ’ ) ;
set ( gcf , ’ Out er Pos i t i on ’ , [ 0 0 1 1 ] ) ;
end
Code block: SLwithoutIns.m:
B.1. HEAT LOSSES FROM SURFACE LINE 135
%% Example 10. 1 Cal c ul at i ons of Heat Los s es from Sur f ace Li nes
% 4 i nc he s N−80 pi pe at r at e of 229 B/D.
% We are l ooki ng f o r the s ol ut i on at steady s t at e c ondi t i ons
% f or pi pe i s not−i ns ul at e d
% wr i t t en 2009 , modi f i ed date 2011 , May 9th
% Fidan , S . ,
%%
cl c ; cl ear al l ; cl ose al l ;
TsteamWOutIns = 550; % F
TaverageWOutIns = 60; % F
PipeLengthWOutIns = 100; % f t
InjTimeWOutIns = 365; % days
AvrWindSpeedWOutIns = 20; % mph
ri nsi deWOutIns = 0. 1478; % f t from Tabl e B. 15
routsi deWOutIns = 0. 1667; % f t
ri nsul atedWOutIns = 0. 4167; % f t
lambdaPipeWOutIns = 600; % Btu/ f t −D−F
lambdaInsWOutIns = 0 . 9 6 ; % Btu/ f t −D−F
hfWOutIns = 48000; % Btu/sq f t −D−F
hpiWOutIns = i nf ; % Btu/sq f t −D−F
hpoWOutIns = 48000; % Btu/sq f t −D−F
% hf c c a l c ul a t i o n based on the Eq . 10. 4
hfcWOutIns = 18∗( AvrWindSpeedWOutIns ˆ 0 . 6 ) . . .
∗( routsi deWOutIns ˆ0. 6) / routsi deWOutIns ;
% Over al l s p e c i f i c thermal r e s i s t a nc e
RhWOutIns = 1/(2∗ pi ) ∗( 1/( hfWOutIns∗ri nsi deWOutIns ) . . .
+ 1/( hpiWOutIns∗ri nsi deWOutIns ) + (1/ lambdaPipeWOutIns ) . . .
∗log ( routsi deWOutIns /ri nsi deWOutIns ) + 1/( ( hfcWOutIns +. . .
htTabl e14 ( TsteamWOutIns , routsi deWOutIns ) ) ∗ . . .
136 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
routsi deWOutIns ) ) ;
di s pl ay ( [ ’ The o v e r a l l s p e c i f i c thermal . . .
r e s i s t a nc e i s c al c ul at e d from Eq . 10. 2 i s =. . .
’ num2str(RhWOutIns) ’ (BTU/ f t −D−ˆF)ˆ−1 ’ ] )
% Heat Los s es
QlsWOutIns = ( TsteamWOutIns − TaverageWOutIns ) . . .
/RhWOutIns ;
di s pl ay ( [ ’ Heat Los s es Ql s = ’ num2str( QlsWOutIns ) . . .
’ Btu/ f t −D’ ] )
% Amount of Heat Lost from the pi pe over a per i od of ti me
QlWOutIns = QlsWOutIns∗InjTimeWOutIns . . .
∗PipeLengthWOutIns ;
di s pl ay ( [ ’ Cumul ati ve Heat Los s es over the per i od . . .
of ti me woth gi ven pi pel engt h Ql =. . .
’ num2str( QlWOutIns ) ’ Btu ’ ] )
h = fi gure ( ’ Col or ’ , [ 0 0 0 ] ) ;
InjTimeplotWOutIns =(1: InjTimeWOutIns ) ’ ;
PlotQlWOutIns =(QlsWOutIns∗PipeLengthWOutIns ) . . .
. ∗ InjTimeplotWOutIns ;
pl otHeatLoss ( PlotQlWOutIns , InjTimeplotWOutIns , . . .
h , QlWOutIns , PipeLengthWOutIns , lambdaInsWOutIns , ’ r ’ ) ;
set ( gcf , ’ Uni ts ’ , ’ normal i zed ’ ) ;
set ( gcf , ’ Out er Pos i t i on ’ , [ 0 0 1 1 ] ) ;
B.2. HEAT LOSSES FROM SEA PART 137
B.2 Heat Losses from Sea Part
(a) Sea Part with insulation. (b) Sea Part without insulation.
Figure B.2: Schematic representation of resistance to heat transfer with temperature
proﬁle.
The following are the MatLab codes for our heat loss calculation starting from sea
surface to sea ﬂoor with insulation and without insulation
%% Sol ut i on f o r Of f s hor e part wi th i ns ul a t i o n mat e r i l as
% We are l ooki ng f o r the s ol ut i on at steady s t at e c ondi t i ons
% wr i t t en 2009 , modi f i ed date 2011 , May 9th
% Fidan , S . ,
%%
cl c ; cl ear al l ; cl ose al l ;
r r i s e r o =18/12/2;
r r i s e r i =0. 60;
r t o = 0. 1458;
r c i = r r i s e r i ;
138 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
rco = r r i s e r o ;
rh= 1;
roWi thIns= r t o ;
r i ns Wi t hI ns = 0. 2292;
r ci Wi t hI ns= r r i s e r i ;
rcoWi thIns= r r i s e r o ;
rwWithIns = rh ;
rEaWithIns = 0;
TAWithIns = 70;
r i ns=r i ns Wi t hI ns ;
T=(TAWithIns −32) /1. 8;
S=35;
al pha1 = SW Di f f us i vi t y (T, S ) ;% i n [mˆ2/ s ]
% 1 [mˆ2/ s ] =38750. 077500155 [ f t ˆ2/ hr ] from
q1=al pha1 ∗38750. 077500155; %[ f t ˆ2/ hr ]
e ps c i = 0 . 9 ; eps t o= 0 . 9 ; e p s r i s e r = 0 . 9 ;
% thermal c onduc t i vi t y of water
%i n [W/m K] ==1 Btu/( h r f t ?F) = 1. 730735 W/( mK ) .
k1 = SW Conducti vi ty (T, S ) ;
%[ Perry ’ s Chemi cal Engi neers ’ Handbook , 7th Edi ti on , Tabl e 1−4]
ke=k1 /1. 730735; % conver s i on f a c t o r
LamdaTub=600/24;
aEWithIns= q1 ; ei nsWi thI ns= 0 . 9 ; eci Wi t hI ns= 0 . 9 ;
lamda EWithIns = ke ; % Btu/ f t −hr−F
l amda i nsWi thIns = [ 0. 166 0. 194 0. 388 0. 499 0. 569 0 . 9 6 ] . / 2 4 ;
InjTi meWi thIns = 21∗24; % hrs
TbWithIns = 600; % F
Pi peLengthWi thIns = 164; % f t
ki ns=l amda i nsWi thIns ;
B.2. HEAT LOSSES FROM SEA PART 139
t=InjTi meWi thIns ;
lamdaPipe = 600/24;
Boltzman = ( 0. 1714 e −8); g =32. 2∗3600∗3600; Tgeo=TAWithIns ; Tf = TbWithIns ;
% Tto = TbWithIns ;
Th=TAWithIns ;
c o l o r s = l i n e s ( length( l amda i nsWi thIns ) ) ;
Col l e c t Dat as I nj I ns =[ ] ;
for i =1: length( l amda i nsWi thIns )
% Step 1 as s i gn random val ue f or dqdz
dqdz=5;
% c a l c ul a t e the ol d T i ns from eq . 16
Ti ns ol d=TAWithIns ;
% h r from eq . 15
h r=hr ( Ti ns ol d , TAWithIns , epsto , eps ci , r c i , r i ns , Boltzman ) ;
% hc from eq . 9
Tanulus=Tan( Ti ns ol d , TAWithIns ) ;%F
Beta Gr=BetaGr ( Tanulus ) ;%R
vi sAn=vi s cos i t yAn ( Tanulus ) ;%cp
% 1 cp = 1488 l b / f t −second = 1488∗3600 l b / f t −hr
vi sAn = visAn ∗1488∗3600;
den=densi tyAn ( Tanulus ) ;%lbm/ f t ˆ3
Gr1=Gr( Ti ns ol d , TAWithIns , visAn , Beta Gr , den , r c i , r i ns , g ) ;
kha=kha1 ( Tanulus ) ;
Pr1=Pr ( kha , vi sAn ) ;
khc=khc1 ( Gr1 , Pr1 , kha ) ;
140 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
h c=hc ( khc , r c i , r i ns ) ;
% from eq . 8
Uto=UtoCal ( h c , h r , r i ns , rto , ki ns ( i ) ) ;
% f ( t ) from Ramey and Wi l l hi t e
ftD = FTIME( Uto , q1 , t , rh , ke ) ;
Tto=TAWithIns+dqdz /(2∗ pi ∗ r t o ∗( h c+h r ))+dqdz /(2∗ pi ∗ ki ns ( i ) ) ∗ . . .
log ( r i ns / r t o ) ;
Ti ns new=Tto−(( dqdz∗log ( r i ns / r t o ) ) /( 2∗ pi ∗ ki ns ( i ) ) ) ;
i t e r =0;
i f abs ( Tins new−Ti ns ol d)<=1
Ti ns new1=Ti ns new ;
Ti ns new=Ti ns new1 ;
el se
while abs ( Tins new−Ti ns ol d)>1
i t e r=i t e r +1;
Ti ns ol d=Ti ns ol d +0. 7∗( abs ( Tins new−Ti ns ol d ) ) ;
Th=Ti ns ol d−dqdz /(2∗ pi ∗ r t o ∗( h c+h r ))+dqdz∗log ( r r i s e r o / . . .
r r i s e r i ) /( 2∗ pi ∗LamdaTub ) ;
dqdz=2∗pi ∗ r t o ∗Uto∗( Tf−Th) ;
h r=hr ( Ti ns ol d , TAWithIns , epsto , eps ci , r c i , r i ns , Boltzman ) ;
% hc from eq . 9
Tanulus=Tan( Ti ns ol d , TAWithIns ) ;%F
Beta Gr=BetaGr ( Tanulus ) ;%R
vi sAn=vi s cos i t yAn ( Tanulus ) ;%cp
% 1 cp = 1488 l b / f t −second = 1488∗3600
B.2. HEAT LOSSES FROM SEA PART 141
vi sAn = visAn ∗1488∗3600;
den=densi tyAn ( Tanulus ) ;%lbm/ f t ˆ3
Gr1=Gr( Ti ns ol d , TAWithIns , visAn , Beta Gr , den , r c i , r i ns , g ) ;
kha=kha1 ( Tanulus ) ;
Pr1=Pr ( kha , vi sAn ) ;
khc=khc1 ( Gr1 , Pr1 , kha ) ;
h c=hc ( khc , r c i , r i ns ) ;
% from eq . 8
Uto=UtoCal ( h c , h r , r i ns , rto , ki ns ( i ) ) ;
% f ( t ) from Ramey and Wi l l hi t e
ftD = FTIME( Uto , q1 , t , rh , ke ) ;
Part1=TAWithIns∗ke ;
Part2=r t o ∗Uto∗ftD∗Tto ;
Part3=ke ;
Part4=r t o ∗Uto∗ftD ;
Th ol d=(Part1+Part2 ) /( Part3+Part4 ) ;
Tto=Ti ns ol d+dqdz∗log ( r i ns / r t o ) /( 2∗ pi ∗ ki ns ( i ) ) ;
Ti ns new=Tto−( r t o ∗Uto ) ∗( Tf−Th ol d )∗ log ( r i ns / r t o ) /( ki ns ( i ) ) ;
end
Ql s=2∗pi ∗ r t o ∗Uto∗( TbWithIns−TAWithIns ) ;
% TOTAL HEAT LOSS
Ql=Ql s ∗Pi peLengthWi thIns ; %BTU
I nj Lengt hpl ot =(1: Pi peLengthWi thIns ) ’ ;
Pl otQl =(Ql s ) . ∗ I nj Lengt hpl ot ;
fi gure ( 1)
142 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
plot ( Pl otQl , I nj Lengt hpl ot , ’ c ol or ’ , c o l o r s ( i , : ) , ’ Li newi dth ’ , 3)
hold on
grid on
set ( gca , ’ XAxi sLocati on ’ , ’ top ’ , ’ YDir ’ , ’ rev ’ )
xlabel ( [ ’ Cumul ati ve Heat Los s es ( i n BTU) f or ’ . . .
num2str( Pi peLengthWi thIns ) ’ f t ’ ] . . .
, ’ Fontname ’ , ’ Ca l i br i ’ , ’ FontSi ze ’ , 16)
ylabel ( ’ Depth ( f t ) ’ )
LambdaValStr = spri ntf ( ’ %0.3G’ , l amda i nsWi thIns ( i ) . / 2 4 ) ;
l a be l s { i } = [ ’ \lambda = ’ num2str( LambdaValStr ) ] ;
legend( l abe l s , 2 , ’ Locati on ’ , ’NE’ )
set ( gcf , ’ Uni ts ’ , ’ normal i zed ’ ) ;
set ( gcf , ’ Out er Pos i t i on ’ , [ 0 0 1 1 ] ) ;
end
end
Code block: SeaPartwithoutIns.m:
%% Sol ut i on f o r Of f s hor e part wi th i ns ul a t i o n mat e r i l as
% We are l ooki ng f o r the s o l ut i on at steady s t at e c ondi t i ons
% wr i t t en 2009 , modi f i ed date 2011 , May 9th , Fidan , S . ,
%%
% Heat l o s s c a l c ul a t i o n f o r wi thout i ns ul a t i o n cas e i s done
cl c ; cl ear al l ; cl ose al l ;
r r i s e r o =18/12/2;% f t
r r i s e r i =0. 60;
r t o = 0. 1458; % f t
r c i = r r i s e r i ; % f t
rco = r r i s e r o ; % f t
rh = 1; % f t
B.2. HEAT LOSSES FROM SEA PART 143
rEaWi thoutIns = 0; % f t
TAWithoutIns = 70; % F
T=(TAWithoutIns −32) /1. 8;
S=35;
al pha1 = SW Di f f us i vi t y (T, S ) ;% i n [mˆ2/ s ]
% 1 [mˆ2/ s ] =38750. 077500155 [ f t ˆ2/ hr ] from
q1=al pha1 ∗38750. 077500155; %[ f t ˆ2/ hr ]
e ps c i = 0 . 9 ;
eps t o = 0 . 9 ;
e p s r i s e r = 0 . 9 ;
% thermal c onduc t i vi t y of water
%i n [W/m K] ==1 Btu/( h r f t ?F) = 1. 730735 W/( mK)
k1 = SW Conducti vi ty (T, S ) ;
%[ Perry ’ s Chemi cal Engi neers ’ Handbook , 7th Edi ti on , Tabl e 1−4]
ke=k1 /1. 730735; % conver s i on f a c t o r
t = 21∗24; % hrs
TbWithoutIns = 600; % F
Pi peLengthWi thoutIns = 164; % 50 meters % f t
lamdaPipe = 600/24;
Boltzman = ( 0. 1714 e −8);
g =32. 2∗3600∗3600;
Tgeo=TAWithoutIns ;
Tf = TbWithoutIns ;
% c a l c ul a t e the ol d T i ns from eq . 16
Tto = TbWithoutIns ;%F
% h r from eq . 15
h r=hr ( Tto , TAWithoutIns , epsto , eps ci , r c i , rto , Boltzman ) ;%btu/hr−f t 2 −F
% hc from eq . 9
Tanulus=Tan( Tto , TAWithoutIns ) ;%F
Beta Gr=BetaGr ( Tanulus ) ;%R
vi sAn=vi s cos i t yAn ( Tanulus ) ;%cp
% 1 cp = 1488 l b / f t −second = 1488∗3600 l b / f t −hr
144 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
vi sAn = visAn ∗1488∗3600;
den=densi tyAn ( Tanulus ) ;%lbm/ f t ˆ3
Gr1=Gr( Tto , TAWithoutIns , visAn , Beta Gr , den , r c i , rto , g ) ;
kha=kha1 ( Tanulus ) ;
Pr1=Pr ( kha , vi sAn ) ;
khc=khc1 ( Gr1 , Pr1 , kha ) ;
h c=hc ( khc , r c i , r t o ) ;
% from eq . 8
Uto=UtoCal ( h c , h r , rh , rco , rto , ke ) ;
% f ( t ) from Ramey and Wi l l hi t e
ftD = FTIME( Uto , q1 , t , rh , ke ) ;
Ql s=2∗pi ∗ r t o ∗Uto∗( TbWithoutIns−TAWithoutIns ) ;
% TOTAL HEAT LOSS
Ql=Ql s ∗Pi peLengthWi thoutIns ; %BTU
I nj Lengt hpl ot =(1: Pi peLengthWi thoutIns ) ’ ;
Pl otQl =(Ql s ) . ∗ I nj Lengt hpl ot ;
fi gure ( 1)
plot ( Pl otQl , I nj Lengt hpl ot , ’ b ’ , ’ Li newi dth ’ , 3)
% Create t i t l e
t i t l e Va l St r = spri ntf ( ’ %0.3G’ , Ql ) ;
t i t l e ( [ ’ Heat Los s es wi t hi n al ong the ’ num2str( Pi peLengthWi thoutIns ) . . .
’ f t i s = ’ t i t l e Va l St r ’ i n Btu ’ ] , . . .
’ FontSi ze ’ , 14 , ’ Col or ’ , ’ k ’ ) ;
hold on
grid on
set ( gca , ’ XAxi sLocati on ’ , ’ top ’ , ’ YDir ’ , ’ rev ’ )
xlabel ( ’ Cumul ati ve Heat Los s es i n BTU’ , ’ Fontname ’ , . . .
’ Cal i br i ’ , ’ FontSi ze ’ , 16)
ylabel ( ’ Depth ( f t ) ’ )
set ( gcf , ’ Uni ts ’ , ’ normal i zed ’ ) ;
set ( gcf , ’ Out er Pos i t i on ’ , [ 0 0 1 1 ] ) ;
B.3. HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 145
B.3 Heat Losses from Sea Floor to Reservoir
(a) Sea Floor to Reservoir with insulation. (b) Sea Floor to Reservoir without insulation.
Figure B.3: Schematic representation of resistance to heat transfer sea ﬂoor to reser
voir.
The following are the MatLab codes for our heat loss calculation starting from sea
ﬂoor to reservoir with insulation and without insulation
%% Example 10. 2 Cal c ul at i ons of Heat Los s es from Sur f ace Li nes
% 3. 5 i n . tubi ng s e t on a packer i n 9 5/8 i n , 53. 5 lbm/ f t N−80 cas i ng
% The annul us c ont at i ns a stagnant gas at zer o gauge pr e s s ur e at wel l head
% and cas i ng i s cemented to s ur f ac e i n a 12−i n hol e .
% We are l ooki ng f o r the s ol ut i on at steady s t at e c ondi t i ons
% wr i t t en 2009 , modi f i ed date 2011 , May 9th
% Fidan , S . ,
%%
cl c ; cl ear al l ; cl ose al l ;
146 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
roWi thIns = 0. 1458; % f t
r i ns Wi t hI ns = 0. 2292; % f t
r ci Wi t hI ns = 0. 3556; % f t
rcoWi thIns = 0. 4010; % f t
rwWithIns = 0. 5000; % f t
rEaWithIns = 0; % f t
aEWithIns = 0 . 9 6 ;
ei nsWi thI ns = 0 . 9 ;
eci Wi t hI ns = 0 . 9 ;
lamda EWithIns = 24; % Btu/ f t −hr−F
lamda cemWithIns = 12; % Btu/ f t −hr−F
l amda i nsWi thIns = [ 0. 166 0. 194 0. 388 0. 499 0. 569 0 . 9 6 ] ;
InjTi meWi thIns = 21; % days
TAWithIns = 100; % F
TbWithIns = 600; % F
Pi peLengthWi thIns = 1000; % f t
c o l o r s = l i n e s ( length( l amda i nsWi thIns ) ) ;
Col l e c t Dat as I nj I ns =[ ] ;
for i =1: length( l amda i nsWi thIns )
% STEP1: I n i t i a l Assumption of the Total Thermal Res i s t ance
Rh=(log ( r i ns Wi t hI ns /roWi thIns )/ l amda i nsWi thIns ( i ) ) / pi ;
% STEP2: Cal cul at e f ( tD) at t
tD=aEWithIns∗InjTi meWi thIns /( rwWithIns ˆ2) ;
Rh2=Rh;
c=2∗pi ∗Rh2∗lamda EWithIns ;
f=f t ( c , tD) ;
ftD=f ;
% STEP3: Cal cul at e Tci from Eq . B68
Up=0;
B.3. HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 147
Ucem=log ( rwWithIns / rcoWi thIns )/ lamda cemWithIns ;
i f rEaWithIns >0
Uea=log ( rEaWithIns /rwWithIns )/ lamda EWithIns ;
el se
Uea=0;
end
Uf=ftD/lamda EWithIns ;
Tci=TAWithIns+((TbWithIns−TAWithIns ) /( 2∗ pi ∗Rh ) ) . . .
∗(Up+Ucem+Uea+Uf ) ;
% STEP4: Cal cul at e Ti ns from Eq . B. 70
Ui =0; Upi =0;Upw=0;Upo=0;
Ui ns=log ( r i ns Wi t hI ns /roWi thIns )/ l amda i nsWi thIns ( i ) ;
Ti ns=TbWithIns −((TbWithIns−TAWithIns ) /( 2∗ pi ∗Rh ) ) . . .
∗( Ui+Upi+Upw+Upo+Ui ns ) ;
%STEP5: Cal cul at e hcan from Eq . B. 63 through B. 66
Tan=(Ti ns+Tci ) /2;
den=0. 076∗((460+60)/(460+Tan ) ) ;
% From FigB. 41 vi s =2.54∗10ˆ( −5)∗T+0.0164
vi s =2.54∗10ˆ( −5)∗Tan+0. 0164;
% l amda a =0. 45;
l amda a = ( 2. 8 e −004)∗Tan+0. 312;
Ban=1/(460+Tan ) ;
g=1; gc =1;
Ngr=(gc /g ) ∗7. 12∗10ˆ7∗( rci Wi thI ns −r i ns Wi t hI ns )ˆ3∗ g ∗ . . .
denˆ2∗Ban∗( Tins−Tci ) /( gc ∗ vi s ˆ2) ;
% Npr ˆ0. 4=0. 92
l amda aa =(0. 049∗ l amda a∗Ngr ˆ0. 333∗0. 92) ;
F=((460+Ti ns )ˆ2+(460+Tci )ˆ2)∗(920+Ti ns+Tci ) ;
hcan=4. 11∗10ˆ( −8)∗((1/ ei nsWi thI ns )+( r i ns Wi t hI ns / r ci Wi t hI ns ) ∗ . . .
(1/ eci Wi thIns −1))ˆ( −1)∗F+l amda aa /( r i ns Wi t hI ns ∗ . . .
log ( r ci Wi t hI ns / r i ns Wi t hI ns ) ) ;
% STEP6: Cal cul at e Rh us i ng Eq . 10. 6
148 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
Ucan=1/(hcan∗ r i ns Wi t hI ns ) ;
Rhc=1/(2∗pi ) ∗( Ui+Upi+Upw+Upo+Ui ns+Ucan+Up+Ucem+Uea+Uf ) ;
counter =0;
while abs ( Rhc−Rh) > 1e−4
counter=counter +1;
Rh=Rhc ;
% STEP2: Cal cul at e f ( tD) at t
tD=aEWithIns∗InjTi meWi thIns /( rwWithIns ˆ2) ;
Rh2=Rh;
c=2∗pi ∗Rh2∗lamda EWithIns ;
f=f t ( c , tD) ;
ftD=f ;
% STEP3: Cal cul at e Tci from Eq . B68
Up=0;
Ucem=log ( rwWithIns / rcoWi thIns )/ lamda cemWithIns ;
i f rEaWithIns >0
Uea=log ( rEaWithIns /rwWithIns )/ lamda EWithIns ;
el se
Uea=0;
end
Uf=ftD/lamda EWithIns ;
Tci=TAWithIns+((TbWithIns−TAWithIns ) /( 2∗ pi ∗Rh ) ) . . .
∗(Up+Ucem+Uea+Uf ) ;
% STEP4: Cal cul at e Ti ns from Eq . B. 70
Ui =0; Upi =0; Upw=0; Upo=0;
Ui ns=log ( r i ns Wi t hI ns /roWi thIns )/ l amda i nsWi thIns ( i ) ;
Ti ns=TbWithIns −((TbWithIns−TAWithIns ) /( 2∗ pi ∗Rh ) ) ∗ . . .
( Ui+Upi+Upw+Upo+Ui ns ) ;
%STEP5: Cal cul at e hcan from Eq . B. 63 through B. 66
Tan=(Ti ns+Tci ) /2;
den=0. 076∗((460+60)/(460+Tan ) ) ;
% From FigB. 41 vi s =2.54∗10ˆ( −5)∗T+0.0164
B.3. HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 149
vi s =2.54∗10ˆ( −5)∗Tan+0. 0164;
l amda a = ( 2. 8 e −004)∗Tan+0. 312;
Ban=1/(460+Tan ) ;
g=1; gc =1;
Ngr=(gc /g ) ∗7. 12∗10ˆ7∗( rci Wi thI ns −r i ns Wi t hI ns ) ˆ 3 ∗ . . .
g∗denˆ2∗Ban∗( Tins−Tci ) /( gc ∗ vi s ˆ2) ;
l amda aa =(0. 049∗ l amda a∗Ngr ˆ0. 333∗0. 92) ;
F=((460+Ti ns )ˆ2+(460+Tci )ˆ2)∗(920+Ti ns+Tci ) ;
hcan=4. 11∗10ˆ( −8)∗((1/ ei nsWi thI ns ) +. . .
( r i ns Wi t hI ns / r ci Wi t hI ns ) ∗ . . .
(1/ eci Wi thIns −1))ˆ( −1)∗F+l amda aa /( r i ns Wi t hI ns ∗ . . .
log ( r ci Wi t hI ns / r i ns Wi t hI ns ) ) ;
% STEP6: Cal cul at e Rh us i ng Eq . 10. 6
Ucan=1/(hcan∗ r i ns Wi t hI ns ) ;
Rhc=1/(2∗pi ) ∗( Ui+Upi+Upw+Upo+Ui ns+Ucan+Up+Ucem+Uea+Uf ) ;
end
Rh=Rhc ;
% HEAT LOSS PER UNIT LENGTH
Ql s=(TbWithIns−TAWithIns )/Rh;
% TOTAL HEAT LOSS
Ql=Ql s ∗Pi peLengthWi thIns ;
Col l e c t Dat as I nj I ns =[ Col l e c t Dat as I nj I ns ; [ Rh Ql s Ql ] ] ;
I nj Lengt hpl ot =(1: Pi peLengthWi thIns ) ’ ;
Pl otQl =(Ql s ) . ∗ I nj Lengt hpl ot ;
fi gure ( 1)
plot ( Pl otQl , I nj Lengt hpl ot , ’ c ol or ’ , c o l o r s ( i , : ) , ’ Li newi dth ’ , 3)
hold on
grid on
set ( gca , ’ XAxi sLocati on ’ , ’ top ’ , ’ YDir ’ , ’ rev ’ )
xlabel ( [ ’ Cumul ati ve Heat Los s es ( i n BTU) f or ’ . . .
num2str( Pi peLengthWi thIns ) ’ f t ’ ] , ’ Fontname ’ , ’ Cal i br i ’ , . . .
’ FontSi ze ’ , 16)
150 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
ylabel ( ’ Depth ( f t ) ’ )
LambdaValStr = spri ntf ( ’ %0.3G’ , l amda i nsWi thIns ( i ) . / 2 4 ) ;
l a be l s { i } = [ ’ \lambda = ’ num2str( LambdaValStr ) ] ;
legend( l abe l s , 2 , ’ Locati on ’ , ’NE’ )
set ( gcf , ’ Uni ts ’ , ’ normal i zed ’ ) ;
set ( gcf , ’ Out er Pos i t i on ’ , [ 0 0 1 1 ] ) ;
end
Code block: InjWellwithoutIns.m:
%% Example 10. 2 Cal c ul at i ons of Heat Los s es from Wel l bore
% 3. 5 i n . tubi ng s e t on a packer i n 9 5/8 i n , 53. 5 lbm/ f t N−80 cas i ng
% The annul us c ont at i ns a stagnant gas at zer o gauge pr e s s ur e at wel l head
% and cas i ng i s cemented to s ur f ac e i n a 12−i n hol e .
% We are l ooki ng f o r the s o l ut i on at steady s t at e c ondi t i ons
% wr i t t en 2009 , modi f i ed date 2011 , May 9th
% Fidan , S . ,
%%
cl c ; cl ear al l ; cl ose al l ;
roWi thoutIns = 0. 1458; % f t
r ci Wi t hout I ns = 0. 3556; % f t
rcoWi thoutIns = 0. 4010; % f t
rwWi thoutIns = 0. 5000; % f t
rEaWi thoutIns = 0; % f t
aEWithoutIns = 0 . 9 6 ;
ei nsWi thoutI ns = 0 . 9 ;
eci Wi t hout I ns = 0 . 9 ;
l amda EWi thoutIns = 24; % Btu/ f t −hr−F
lamda cemWithoutIns = 12; % Btu/ f t −hr−F
Inj Ti meWi thoutIns = 21; % days
TAWithoutIns = 100; % F
TbWithoutIns = 600; % F
Pi peLengthWi thoutIns = 1000; % f t
B.3. HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 151
lamdaPipe =600;
% Step 1 as s i gn random val ue f o r dqdz
dqdz=randperm( 10 0) ;
dqdz=dqdz ( 1 ) ;
% c a l c ul a t e the ol d T i ns from eq . 16
% Step 2: Tci = Geothermal Temperature
Tc i ol d=TAWithoutIns ;
% Step 3: h r from eq . 15 , hc from eq . 9 and
% Uto from eq . 8 and f ( t ) from Ramey
% h r from eq . 15
h r=hr ( Ti ns ol d , Tci ol d , eps i ns , eps ci , r c i , r i ns , Boltzman ) ;
% hc from eq . 9
Tanulus=Tan( Ti ns ol d , Tc i ol d ) ;
Beta Gr=BetaGr ( Tanulus ) ;
vi sAn=vi s cos i t yAn ( Tanulus ) ;
den=densi tyAn ( Tanulus ) ;
Gr1=Gr( Ti ns ol d , Tci ol d , visAn , Beta Gr , den , r c i , r i ns , g ) ;
kha=kha1 ( Tanulus ) ;
Pr1=Pr ( ) ;%kha , visAn
khc=khc1 ( Gr1 , Pr1 , kha ) ;
h c=hc ( khc , r c i , r i ns ) ;
% from eq . 8
Uto=UtoCal ( h c , h r , rh , r i ns , rco , rto , ki ns , kcement ) ;
% f ( t ) from Ramey and Wi l l hi t e
ftD = FTIME( Uto , q1 , t , rh , ke ) ;
% s t ep 4: c a l c ul a t e the ol d Th
Th ol d=( r t o ∗Uto∗ftD∗Tf+kcement ∗Tgeo ) /( r t o ∗Uto∗ftD+kcement ) ;
% s t ep 5: c a l c ul a t e the new Tci
152 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
Tci new=Th ol d+(( r t o ∗Uto∗log ( rh/ rco ) ) /( kcement ) ) ∗( Tto−Th ol d ) ;
i t e r =0;
i f abs ( Tci new−Tc i ol d)<=1
Tci new1=Tci new ;
Tci new=Tci new1 ;
el se
while abs ( Tci new−Tc i ol d)>=1e−4
i t e r=i t e r +1;
Tc i ol d=Tc i ol d +0. 7∗( abs ( Tci new−Tc i ol d ) ) ;
Part1=Tc i ol d ∗kcement ;
Part2=r t o ∗Uto∗log ( rh/ rco )∗ Tto ;
Part3=kcement ;
Part4=r t o ∗Uto∗log ( rh/ rco ) ;
Th=(Part1−Part2 ) /( Part3−Part4 ) ;
dqdz=2∗pi ∗ r t o ∗Uto∗( Tf−Th) ;
Ti ns ol d=Tto−(( dqdz∗log ( r i ns / r t o ) ) /( 2∗ pi ∗ ki ns ) ) ;
% Step 3: h r from eq . 15 , hc from eq . 9 and Uto
%from eq . 8 and f ( t ) from Ramey
% h r from eq . 15
h r=hr ( Ti ns ol d , Tci ol d , eps i ns , eps ci , r c i , r i ns , Boltzman ) ;
% hc from eq . 9
Tanulus=Tan( Ti ns ol d , Tc i ol d ) ;
Beta Gr=BetaGr ( Tanulus ) ;
vi sAn=vi s cos i t yAn ( Tanulus ) ;
den=densi tyAn ( Tanulus ) ;
Gr1=Gr( Ti ns ol d , Tci ol d , visAn , Beta Gr , den , r c i , r i ns , g ) ;
kha=kha1 ( Tanulus ) ;
khc=khc1 ( Gr1 , Pr1 , kha ) ;
h c=hc ( khc , r c i , r i ns ) ;
% from eq . 8
B.3. HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 153
Uto=UtoCal ( h c , h r , rh , r i ns , rco , rto , ki ns , kcement ) ;
% f ( t ) from Ramey and Wi l l hi t e
ftD = FTIME( Uto , q1 , t , rh , ke ) ;
% s t ep 4: c a l c ul a t e the ol d Th
Th ol d=( r t o ∗Uto∗ftD∗Tf+kcement ∗Tgeo ) /( r t o ∗Uto∗ftD+kcement ) ;
% s t ep 5: c a l c ul a t e the new Tci
Tci new=Th ol d+(( r t o ∗Uto∗log ( rh/ rco ) ) /( kcement ) ) ∗( Tto−Th ol d ) ;
end
end
154 APPENDIX B. CODES FOR HEAT LOSS CALCULATIONS
B.4 Table 14 from Prats [41]
The following are the MatLab codes for our heat loss calculation for surface lines
without insulation radiation and free convection number interpolated values using
Table B.14 from Prats[41].
function ht=htTabl e14 (Temp, Di ameter )
% Fidan , S .
% I nt e r pol a t i o n f o r the htc
% Prats pg234 t abl e B. 14
Tabl e14 =[ 50. 9 59. 5 66. 2 74. 4 81. 8 90 107 127 149 174 202 234 269 307 352
48. 7 57. 1 63. 6 71. 5 79 86. 9 104 124 146 171 198 230 265 304 348
46. 3 54. 5 60. 5 68. 4 75. 4 83. 3 100 120 141 166 194 225 260 299 343
44. 2 51. 8 57. 8 65. 3 72. 2 79. 9 96. 5 116 137 162 189 221 256 294 338
42. 2 49. 4 70. 1 62. 4 69. 4 76. 8 93. 1 112 134 158 186 217 252 290 334
41. 0 48. 2 53. 8 61. 0 67. 7 75. 1 91. 9 111 132 156 184 215 250 289 332
39. 4 46. 3 51. 6 58. 8 65. 3 72. 7 88. 8 108 129 153 180 212 247 286 3 2 9 ] ;
DimPipeIns =[ 0. 5 1 2 4 8 12 2 4 ] ;
SurfTemp=[130 180 230 280 330 380 480 580 680 780 880 980 1080 1180 1280] ;
ht=interp2 ( SurfTemp , DimPipeIns , Table14 , Temp, Di ameter ) ;
return
B.5. F(TD) CALCULATION ALSO KNOWN AS RAMEY[?] 155
B.5 f(tD) calculation also known as Ramey[42]
The following are the MatLab codes for our heat loss calculation for sea ﬂoor to reser
voir with/without insulation interpolated values using Table 10.1 from Prats[41].
function f=f t ( r , tD)
% Fidan , S . ,
% I nt e r po l a t i o n of the val ue of f ( t ) i n unsteady conducti on
A=[ . 311 . 312 . 313 . 313 . 314 . 316 . 318 . 323 . 330 . 345 . 373 . 396 . 417 . 433 . 438 . 445
. 421 . 422 . 423 . 423 . 424 . 427 . 430 . 439 . 452 . 473 . 511 . 538 . 568 . 572 . 578 . 588
. 614 . 615 . 616 . 617 . 619 . 623 . 629 . 644 . 666 . 698 . 745 . 772 . 790 . 802 . 806 . 811
. 800 . 801 . 802 . 803 . 806 . 811 . 820 . 842 . 872 . 910 . 958 . 984 1. 00 1. 01 1. 01 1. 02
1. 00 1. 01 1. 02 1. 02 1. 03 1. 04 1. 05 1. 08 1. 11 1. 15 1. 20 1. 22 1. 24 1. 24 1. 24 1. 25
1. 34 1. 35 1. 36 1. 37 1. 37 1. 38 1. 40 1. 44 1. 48 1. 52 1. 56 1. 57 1. 58 1. 59 1. 59 1. 59
1. 63 1. 64 1. 65 1. 66 1. 66 1. 67 1. 69 1. 73 1. 77 1. 81 1. 84 1. 86 1. 86 1. 87 1. 87 1. 88
1. 94 1. 95 1. 96 1. 97 1. 97 1. 99 2. 00 2. 05 2. 09 2. 12 2. 15 2. 16 2. 16 2. 17 2. 17 2. 17
2. 37 2. 38 2. 39 2. 39 2. 40 2. 42 2. 44 2. 48 2. 51 2. 54 2. 56 2. 57 2. 57 2. 57 2. 58 2. 58
2. 71 2. 72 2. 73 2. 73 2. 74 2. 75 2. 77 2. 81 2. 84 2. 86 2. 88 2. 89 2. 89 2. 89 2. 89 2 . 9 0 ] ;
t =[ . 1 . 2 . 5 1 2 5 10 20 50 1 0 0 ] ;
R=[1000 500 100 50 20 10 5 2 1 . 5 . 2 . 1 . 05 . 02 . 01 0 ] ;
i f tD>100
f =0.5∗ log ( tD) +0. 403;
el se
f=interp2 (R, t , A, r , tD) ;
end
return
Appendix C
Results for Diﬀerent Insulation
Materials
UsingWhiteAerogelλ
WA
= 0.0081BTU/(ft −hr −
o
F)
Figure C.1: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft) , 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
156
157
Figure C.2: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
Figure C.3: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
158 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.4: Heat loss distribution for diﬀerent injection temperature vs depth (ft), 1
year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
UsingFiberglassλ
FG
= 0.0162BTU/(ft −hr −
o
F)
Figure C.5: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
159
Figure C.6: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
Figure C.7: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
160 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.8: Heat loss distribution for diﬀerent injection temperature vs depth (ft), 1
year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
UsingCarbonFiberλ
CF
= 0.0208BTU/(ft −hr −
o
F)
Figure C.9: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
161
Figure C.10: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
Figure C.11: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
162 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.12: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
UsingThermolasticInsulationλ
TI
= 0.0237BTU/(ft −hr −
o
F)
163
Figure C.13: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic
insulation.
Figure C.14: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic insulation.
164 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.15: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic insulation.
Figure C.16: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic insulation.
UsingCalciumSilicateInsulationλ
CaSil
= 0.04BTU/(ft −hr −
o
F)
165
Figure C.17: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
Figure C.18: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
166 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.19: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
Figure C.20: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
167
UsingWhiteAerogelλ
WA
= 0.0081BTU/(ft −hr −
o
F)
Figure C.21: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
Figure C.22: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
168 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.23: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
Figure C.24: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with white aerogel.
UsingFiberglassλ
FG
= 0.0162BTU/(ft −hr −
o
F)
169
Figure C.25: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
Figure C.26: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
170 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.27: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
Figure C.28: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with ﬁber glass.
171
UsingCarbonFiberλ
CF
= 0.0208BTU/(ft −hr −
o
F)
Figure C.29: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
Figure C.30: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
172 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.31: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
Figure C.32: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with carbon ﬁber.
UsingThermolasticInsulationλ
TI
= 0.0237BTU/(ft −hr −
o
F)
173
Figure C.33: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic
insulation.
Figure C.34: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic insulation.
174 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.35: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic insulation.
Figure C.36: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with thermolastic insulation.
UsingCalciumSilicateInsulationλ
CaSil
= 0.04BTU/(ft −hr −
o
F)
175
Figure C.37: Steam temperature distribution for diﬀerent injection temperature vs
depth (ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
Figure C.38: Steam pressure distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
176 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS
Figure C.39: Steam quality distribution for diﬀerent injection temperature vs depth
(ft), 1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
Figure C.40: Heat loss distribution for diﬀerent injection temperature vs depth (ft),
1 year, T
m
= 122
o
F and injection rate 4850 lbm/hr with calcium silicate.
Bibliography
[1] Schematic view of one of the heat transfer mechanism: Convection.
http://www.ic.sunysb.edu/Class/phy141md/doku.php?id=phy141:lectures:35.
[2] Schematic view of the oﬀshore platform. http://www.consrv.ca.gov/dog/Picture
awell/PublishingImages /OFFSHOREPLATFORM2.jp.
[3] Two phase ﬂow correlations : Modiﬁed Beggs and Brill two phase correlations.
www.fekete.com/software/rta/media/webhelp/ctecorrelations.htm.
[4] Two phase ﬂow correlations with pressure loss calculation : Mod
iﬁed Beggs and Brill two phase correlations with Pressure Loss.
www.fekete.com/software/feketeharmony/media/webhelp/HTML
Files/ReferenceMaterial/CalculationsandCorrelations/PressureLoss
Calculations.htm.
[5] Abu AlSoof Nimat B AlNajjar, Hazim S.H. Alternative ﬂowpattern maps can
improve pressuredrop calculations of the aziz et al. multiphaseﬂow correlation.
SPE Production Engineering, 4(3), 1989.
[6] Chen ChouMing Ayres, Paul S. and Burton D. Ziels. Vacuum
insulated steam injection tubing, united states patent 4512721.
http://www.patentgenius.com/patent/4512721.html, 1982.
177
178 BIBLIOGRAPHY
[7] H.D. Beggs and J.P. Brill. A study of twophase ﬂow in inclined pipes. Journal
of Petroleum Technology, 25(5):607–617, 1973.
[8] Kraus Allan D Bejan, Adrian. Heat Transfer Handbook. John Wiley & Sons,
2003.
[9] Dominuque Monfrin Bernard M. Couderc, J.F. Verpeaux and Lisette H. Quettier.
Emeraude vapeur: A steam pilot in an oﬀshore environment. Paper SPE 16723
presented at the SPE Annual Technical Conference and Exhibition held in Dallas,
TX, 2730 September, 1987.
[10] Malcolm Alister Grant & Paul F Bixley. Geothermal Reservoir Engineering.
Academic Press, second edition, 2011.
[11] W.B. Bleakley. Here are case histories of two thermal projects. The Oil and Gas
Journal, pages 123–130, 1964.
[12] Dr. Ove Bratland. The ﬂow assurance site multiphase ﬂow assurance.
http://www.drbratland.com/PipeFlow2/chapter1.html, 2010.
[13] J. P. Brill and H. D. Beggs. TwoPhase Flow in Pipes.
[14] J. P.. Brill and H. Mukherjee. Multiphase Flow in Wells. Society of Petroleum
Engineers, 1999.
[15] G.W. Runberg A.J. Cornelius B.T. Willman, V.V. Valleroy and L.W. Powers.
Laboratory studies of oil recovery by steam injection. Journal of Petroleum
Technology, 13(7):681–690, 1961.
[16] Jr. D.H. Gibson S.W. Gosch P.D. Pattillo J.W. Sharp C.E. Taylor D.W. Brad
ford, D.G. Fritchie. Marlin failure analysis and redesign; part 1, description of
BIBLIOGRAPHY 179
failure. Paper SPE 74528 presented at the IADC/SPE Drilling Conference, 2628
February 2002, Dallas, Texas.
[17] Robert C. Earlougher Jr. Some practical considerations in the design of steam
injection wells. Journal of Petroleum Technology, 21(1):79–86, 1969.
[18] S.M. Farouq Ali. A comprehensive wellbore stream/water ﬂow model for steam
injection and geothermal applications. SPE Journal, 21(5):527–534, 1981.
[19] J. P. Fontanilla. A Mathematical Model for the Prediction of Wellbore Heat Loss
and Pressure Drop in Steam Injection Wells. Master’s thesis, Department of
Chemical Engineering, University of Calgary, 1980.
[20] Jerry P. Fontanilla and Khalid Aziz. Prediction of bottomhole conditions for wet
steam injection wells. Journal of Canadian Petroleum Technology, 21(2):82–88,
1982.
[21] L.J. Durlofsky K. Aziz L.R. Diaz B. Alkaya G. Oddie H. Shi, J.A. Holmes. Drift
ﬂux modeling of multiphase ﬂow in wellbores. SPE Annual Technical Conference
and Exhibition, 58 October 2003, Denver, Colorado, 2003.
[22] Alton R. Hagedorn and Kermit E. Brown. Experimental study of pressure gradi
ents occurring during continuous twophase ﬂow in smalldiameter vertical con
duits. Journal of Petroleum Technology, 17(4):475–484, 1965.
[23] A.R. Hasan and C. S. Kabir. Fluid Flow and Heat Transfer in Wellbores. Society
of Petroleum Engineers, ﬁrst edition, 2002.
[24] M. Holmgren. X steam for matlab. http://www.xeng.com, 2006.
[25] P.H. Holst and D.L. Flock. Wellbore behaviour during saturated steam injection.
Journal of Canadian Petroleum Technology, 5(4):184–193, 1966.
180 BIBLIOGRAPHY
[26] Hans H. A. Huygen and J. L. Huitt. Wellbore heat losses and casing temperatures
during steam injection. American Petroleum Institute, pages 25–32, 1966.
[27] Jan Dirk Jansen. Modeling and Optimization of Oil and Gas Production Systems.
Advanced Production Engineering Lecture Notes, 2011.
[28] H. Duns Jr. and N. C. J. Ros. Vertical ﬂow of gas and liquid mixtures in wells. 6th
World Petroleum Congress, June 19  26, 1963 , Frankfurt am Main, Germany,
1963.
[29] and Maria Fogarasi Khalid Aziz, George W. Govier. Pressure drop in wells
producing oil and gas. Journal of Canadian Petroleum Technology, 11(3), 1972.
[30] L.W. Lake. Enhanced Oil Recovery. Society of Petroleum Engineers, 1989.
[31] K. Leutwyler. Casing temperature studies in steam injection wells. Journal of
Petroleum Technology, 18(9):1157–1162, 1966.
[32] Miller Mark A. Liang, Zhiyue and Kamy Sepehrnoori. New functional correla
tions for saturated steam properties. Society of Petroleum Engineers, 1992.
[33] D. R. Liles. Twophase ﬂow. Los Alamos Science, pages 26–35.
[34] Cristovao Marques. ThawFront Dynamics of SuperInsulated Wells in Cold
Environments. Master’s thesis, Department of Energy Resources Engineering,
Stanford University, 2007.
[35] A.F. Mills. Basic Heat and Mass Transfer. Prentice Hall, second edition, 1999.
[36] Syed M. Zubair Mostafa H. Sharqawy, John H. Lienhard Va. Thermophysical
properties of seawater: a review of existing correlations and data. Desalination
and Water Treatment, 2009.
BIBLIOGRAPHY 181
[37] Brill James P. Mukherjee, Hemanta. Empirical equations to predict ﬂow patterns
in twophase inclined ﬂow. International Journal of Multiphase Flow, 11(3):299–
315, 1985.
[38] E.F. Pacheco and S.M. Farouq Ali. Wellbore heat losses and pressure drop in
steam injection. Journal of Petroleum Technology, 24(2):139–144, 1972.
[39] Palmer C.M. Brill J.P. Payne, G.A. and H.D. Beggs. Evaluation of inclinedpipe,
twophase liquid holdup and pressureloss correlation using experimental data.
Journal of Petroleum Technology, 31(9):1198–1208, 1979.
[40] Palmer C.M. Brill J.P. Payne, G.A. and H.D. Beggs. A study of multiphase ﬂow
behavior in vertical wells. SPE Production Engineering, 3(2):263–272, 1988.
[41] M. Prats. Thermal Recovery. Society of Petroleum Engineers, 1986.
[42] H.J. Ramey JR. Wellbore heat transmission. Journal of Petroleum Technology,
14(4):436–440, 1962.
[43] D.H. Gibson S.W. Gosch P.D. Pattillo R.C. Ellis, D.G. Fritchie Jr. Marlin failure
analysis and redesign: Part 2  redesign. Paper SPE 74529 presented at the
IADC/SPE Drilling Conference, 2628 February 2002, Dallas, Texas.
[44] P. S. Sarathi and D. K. Olsen. Petroleum Reservoir Simulation. Applied Science
Publishers, 1992.
[45] A. Satter. Heat losses during ﬂow of steam down a wellbore. Journal of Petroleum
Technology, 17(7):845–851, 1965.
[46] O. Shoham. Mechanistic Modeling of GasLiquid twophase Flow in Pipes. So
ciety of Petroleum Engineers, ﬁrst edition, 2005.
182 BIBLIOGRAPHY
[47] Smith D.D. Squier, D.P. and E.L. Dougherty. Calculated temperature behavior
of hotwater injection wells. Journal of Petroleum Technology, 14(4):427–435,
1962.
[48] P.D. Pattillo J.W. Sharp S.W. Gosch, D.J. Horne and P.C. Shah. Marlin failure
analysis and redesign; part 3, vit completion with realtime monitoring. Paper
SPE 74530 presented at the IADC/SPE Drilling Conference, 2628 February
2002, Dallas, Texas.
[49] G.P. Willhite. Overall heat transfer coeﬃcients in steam and hot water injection
wells. Journal of Petroleum Technology, 19(5):607–615, 1967.
[50] YuShu Wu and Karsten Pruess. An analytical solution for wellbore heat trans
mission in layered formations (includes associated papers 23410 and 23411 ).
SPE Reservoir Engineering, 5(4):531–538, 1990.
c Copyright by Sel¸uk Fidan 2011 c All Rights Reserved
ii
I certify that I have read this thesis and that in my opinion it is fully adequate, in scope and in quality, as partial fulﬁllment of the degree of Master of Science in Energy Resources Engineering.
Prof. Anthony R. Kovscek (Principal Adviser)
I certify that I have read this thesis and that in my opinion it is fully adequate, in scope and in quality, as partial fulﬁllment of the degree of Master of Science in Energy Resources Engineering.
Dr. Louis Castanier
iii
iv
Accurate predictions of heat loss. Even today the topic is important for practical application of steam injection. Most thermal reservoir simulators today do not yet take into account heat losses and pressure drops along the wellbore. For deeper injection wells and injection wells in oﬀshore environments. 2) white v . The equations describing mass and heat ﬂow are solved in discretized wellbore framework. Steam quality. are used and results are compared. however.Abstract In the oil industry. wellbore heat loss is often signiﬁcant. In the literature. Six insulation materials are examined: 1) black aerogel. Steam properties are incorporated directly. the problem of wellbore heat loss during hot ﬂuid injection is classical. Several twophase ﬂow correlations for injection tubing. The main goal of this study is to investigate heat losses along the wellbore during steam injection in both onshore and oﬀshore environments. The Fontanilla and Aziz approach is used in this study. 14]. steam temperature. and heat loss values with and without insulation are calculated. temperature distributions and pressure proﬁle are essential for modeling steam injection wells. with an improvement in the application of twophase ﬂow correlations and the determination of several input parameters. Neglecting these items may be acceptable for shallow reservoirs. The calculated steam temperature and steam pressure agree well with the ﬁeld data using the Beggs and Brill model [13. it is shown that the Fontanilla and Aziz model [20] yields results in good agreement with ﬁeld data. steam pressure.
vi . so are the amounts of heat loss. Compared with the case of just steam injection. the steam temperature values are smaller. that enables other users to change the input parameters and visualize the results without going into the details of the calculations. and heat loss calculation are conducted for steam injection with noncondensible gas (N2 ). steam temperature. 5) thermolastic insulation and 6) calcium silicate.aerogel. Steam quality. 4) carbon ﬁber. a novel approach is introduced for adding noncondensible gas to steam to increase the injection pressure without increasing the steam temperature. A Matlab Graphical User Interface (GU I) is developed. steam pressure. To our knowledge. 3) ﬁberglass. Aerogel insulations present the opportunity to create a superinsulated tubing that overcomes many limitations of current steam injectors. no one has predicted the result of noncondensible gas addition on steam injectors. In this work. Additional partial pressure is obtained by adding N2 to the system.
this success was not only one person’ success it was the success of the several people in my life and I will brieﬂy talk about those people here. My mother is the highest priority person in my life not only she raised us with devoting her entire life to her children but also lack of opportunity she had not to allowed to get educated. Mustafa Onur who has great impact on my life in terms of his knowledge. Dr. I would like to thank both to Turkish National Petroleum Cooperation (T. Kovscek for his time.Acknowledgments First and foremost. He always gave his students courage to excel their skills and motivated them to work vii .A. It came true. My adviser in Turkey Prof. support and trust on me. Khalid Aziz to provide one of his Master students Fontanilla’s thesis. Jan Dirk Jensen to teach us ”Design and analysis of production systems for oil and gas reservoirs” to understand multiphase ﬂow concept better. I would also like to thank to visiting Prof. He is a unique person both in Academic world and personal world. support.P. It was one of the biggest dream I had since second year of undergraduate to come to Stanford and study there.O) and SUPRIA Aﬃliates for their support. My special thanks go to Prof. thank you ’Anne’ (means mom in Turkish). Louis Castanier with whom I had helpful discussions in the early and late stage of this project. Anthony R. I would like to express my sincere gratitude to my adviser Prof. Thank you Tony! My thanks also go to Dr. she dedicated herself to her children to get all of them (4 sons) educated and she achieved this. guidance and his unbelievable patience. Without his support I would not be able to ﬁnish this work.
hard. Amar Alshehri. Mehrdad Honarkhah. Wenjuan Lin. not only several days I stayed at school she never became angry. Elnur Aliyev. Rustem Zaydullin. Now my friends . many thanks to him and as well as my supervisor Dr. It was my fortune to be one of the member of the such a great team SUPRIA that I have learned a lot and found my women of dreams and got married. many thanks to him. he gave me an opportunity to work with him and learn from his experiences. thank you little angel. and energy to our life. is the softest part of my heart and inspiration of my life. Alireza Iranshahr. I enable to ﬁnish this work! Thank you.). and more. joy. My wife. Obi Isebor: I enjoyed studying with you guys during my stay at Stanford. viii . My adviser Dr. Mehmet Parlar at Schlumberger during my internship last summer. CANIM! And my little baby daughter Su Lin Fidan who brought fun. She always supported me. because of him I am here. but also she showed her love to me every seconds of our life. Because of her love. Rajesh Chanpura.
If you can trust yourself when all men doubt you. being hated. Or. Or walk with kings . And lose. If you can make one heap of all your winnings And risk it on one turn of pitchandtoss. And yet don’t look too good.you’ll be a Man my son! Rudyard Kipling ix . If you can think . If you can bear to hear the truth you’ve spoken Twisted by knaves to make a trap for fools. But make allowance for their doubting too. If you can meet with triumph and disaster And treat those two imposters just the same. If you can talk with crowds and keep your virtue.which is more . And so hold on when there is nothing in you Except the Will which says to them: ”Hold on”. being lied about. And stoop and build ’em up with wornout tools. but none too much. And . Or. If you can force your heart and nerve and sinew To serve your turn long after they are gone.IF If you can keep your head when all about you Are losing theirs and blaming it on you.nor lose the common touch. If neither foes nor loving friends can hurt you.and not make thoughts your aim. If you can ﬁll the unforgiving minute With sixty seconds’ worth of distance run Yours is the Earth and everything that’s in it. don’t give way to hating. Or watch the things you gave your life to broken. If you can wait and not be tired by waiting. nor talk too wise. If all men count with you.and not make dreams your master. and start again at your beginnings And never breath a word about your loss. If you can dream . don’t deal in lies.
P) u u x .Dedicated to my father H¨sn¨ Fidan (R.I.
. . . . .2 2. . .1 Heat Loss from Surface Lines . . . 3. . . . Heat Transfer by Radiation . . .1.2 2. . . 3. 2. . . .1.Contents Abstract Acknowledgments 1 Introduction 1. . . . . . . .3. Heat Transmission Mechanisms and Discussion from Authors . .3. . . . . . . . . . . . .1 Thesis Outline .1. . .1.1 2. . . . . . 3 Model Formulation 3. .1 Heat Transmission Mechanisms . . . . . . .1. . 2. . . . . . . . . . . . . .3 2. .1 2. . Marlin Failure Analysis and Redesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . .3. . . . . . . .1 With/without Insulation . .3 Emeraude Vapeur : A Steam Pilot in an Oﬀshore Environment . . . . . . . . . v vii 1 4 6 6 9 11 11 11 12 12 13 15 15 17 17 2 Literature Review 2. . . . . . . Heat Transmission Discussion from Authors . . . . .3. . . . . xi . . . . . . . .3. .2 Heat Transfer by Conduction . . . . . Heat Transfer by Convection . . . . . . . . . . . . . . . .1 Heat Loss Calculations . . . . .
. .1 Modiﬁed Beggs and Brill Model . . . .1 from Prats . . . . 3.3. .3. 3. . . . . . . . . . . . 3. . . . . . . .1. . .1. xii . . . . . . .2 Flow Pattern Determination . . . . .1. .2 3. 6. . . . .1. . . .3. . . . . . . . 4 Eﬀect of NonCondensable Gas (N2 ) 5 Graphical User Interface (GUI) 6 Results and Comparisons 6. . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . 3. .3. . . . . . . . .3 6. . . . . . . . . . . . . . . . 3. .1 3. . . . . . . . . . . . . .1. . . .3. . . . . . . . . . . .1. .3. Aziz. .4 Oﬀshore Environments .2.3.1 With/without Insulation . . . . . .3 Heat Loss from Sea Floor to Reservoir . . .3. 3. .3 3. Examples without Insulation Materials . . .1 Examples for heat loss calculation . . . Example 10. . . .3 Example 10. . . . . . . . . . . .2 6.2 3. . . . . . . . . . . . 6. . . . . . . . . . . . . . . . Govier and Fogarasi Model . . . . . . . . . . . . . . . . . . . . . . .2 Examples with Insulation Materials . . . . . Frictional Pressure Loss . . . . . . . . . . . Modiﬁcations . . . . . . . . . . . .2 from Prats . .3. . . . . . . .2 Heat Loss from Sea Level to Sea Floor . . . .2 FlowPattern Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . 6. . . . Onshore environments . . .3 Steam Phase behavior calculations . . . . . . Hydrostatic Pressure Diﬀerence . . .1. . . . . . . . . . . . . . .2 6. . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . Example for Oﬀshore . . . .1 3. . . . . .1 6. .1 6. . .2. . . .1. . . . . . . . .1 With/without Insulation . . . . Program Validation . 19 19 22 22 26 27 32 33 36 38 39 39 43 45 49 53 53 54 59 60 63 73 75 77 79 3.1. . . . . . . . .3. . Two Phase Flow Correlations . . . . . . . . . .
. .6. . . . . . . . . . . . . . . . . . . . . . . 103 Future Work . . . .6.5 Determination of the Convection Heat Transfer Coeﬃcient . . . . . . . . . . . . . . . .4. Examples without Insulation Materials . . . . . .4. . . . . .3 Summary . . .3 Evaluation of Heat Loss to the Surrounding . . . . . .2 Examples with Insulation Materials . . . . . . . . . . . 122 A. . . . . . . . . . . . . . . . . . . . . . . . 7 Summary. 79 81 84 84 89 95 95 99 102 Adding NonCondensable Gas (N2 ) in an Onshore environment 6. . . . . . . . . . . . 6. . . . . . .6 Determination of the Radiation Heat Transfer Coeﬃcient . . . . 110 A. . . . . . . . . . . .10 Calculation of the Annulus Fluid Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 Evaluation of the Derrivatives . . . . . . . . . . . . . . . . . . . . . . . . 128 xiii . . . . . . . . . . . . . . . . . . 6. . . . . . . .5. . . . . .1 6. . . .1 7. . 116 A.1 Total Energy Equation . . . . . Examples without Insulation Materials . . . . . . . .6 Adding NonCondensable Gas (N2 ) in an Oﬀshore environment . . . . . . 102 Conclusions . . . .1 6. . . .2 7. .2 Mechanical energy balance or the Extended Bernoulli Equation . . . . . . . . . 122 A. .2 Examples with Insulation Materials . . . . . . . . . . . . . . . . . . . . . . .8 Determination of f(t) . . . . 113 A. . . 125 A. . . . . . . . . . . .6. . . . Examples without Insulation Materials . . . . . . . .4 Determination of the Uto and Tci . . . Conclusions and Future Work 7. . . . .7 Computational Procedure for Uto .5. . .5 Examples with Insulation Materials . . . . . . . . . 127 A. 120 A. . 105 Nomenclature . . . . . . . . . . . .2 6. . 106 A Derivation of the Equations 110 A. . .1 6. . . . . . . . . . . . . . . . . . . . . 115 A. . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . 145 B. . 132 B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 B. . . . . .3 Heat Losses from Sea Floor to Reservoir . . . . . . . . . . . . 155 C Results for Diﬀerent Insulation Materials Bibliography 156 177 xiv . . . . . . .2 Heat Losses from Sea Part . .1 Heat Losses from Surface Line . . . . .4 Table 14 from Prats [41] . 154 B. .B Codes for Heat Loss Calculations 132 B. .5 f(tD) calculation also known as Ramey[42] .
List of Tables
3.1 3.2 6.1 Thermal Conductivity of the materials . . . . . . . . . . . . . . . . . Time Function f (tD ) for the boundary condition model [49]. . . . . . Input parameters from Prats [41] as used for diﬀerent example calculations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 6.3 Radiationnatural convection coeﬃcient of heat transfer. . . . . . . . Field data parameters for ﬁeld data 1 and ﬁeld data 2 [19]. . . . . . . 56 57 63 16 25
xv
xvi
List of Figures
1.1 2.1 2.2 2.3 Schematic view of the objective of our calculations, (retrieved from [2]). Emeraude ﬁeld location and ﬁve spot[9]. . . . . . . . . . . . . . . . . Schematic view of conduction (after [8]). . . . . . . . . . . . . . . . . Heat transfer from a hot surface to air by convection (retrieved from [1]). 2.4 3.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 12 4 7 11
Representation of heat transfer by radiation(after [8]). . . . . . . . . . Schematic representation of resistance to heat transfer with or without temperature proﬁle. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
3.2
Schematic representation of resistance to heat transfer with temperature proﬁle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 21 21
3.3 3.4 3.5
Sea water properties change with temperature and salinity[36]. . . . . Schematic representation of the wellbore. . . . . . . . . . . . . . . . . Schematic representation of resistance to heat transfer sea ﬂoor to reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
22 26 27 29 30 34
3.6 3.7 3.8 3.9
Pressureenthalpy diagram (retrieved from [24]). . . . . . . . . . . . . Gasliquid ﬂowpatterns for vertical pipes (retrieved from [12]). . . . . Vertical downward twophase ﬂow [33]. . . . . . . . . . . . . . . . .
Liquid Holdup and Slippage eﬀect representation (retrieved from[4]).
3.10 Flow Map for the Beggs and Brill Correlation (retrieved from[4]). . . xvii
3.11 Segregated Flow Regime (retrieved from[4]). . . . . . . . . . . . . . . 3.12 Intermittent Flow Regime (retrieved from[4]). . . . . . . . . . . . . . 3.13 Distributed Flow Regime (retrieved from[4]). . . . . . . . . . . . . . . 3.14 Flow Pattern map for Aziz et al. (retrieved from[14]). . . . . . . . . . 5.1 5.2 5.3 5.4 User interface developed GUI for onshore calculations. . . . . . . . . User interface developed GUI for oﬀshore calculations. . . . . . . . . User interface developed GUI for both onshore and oﬀshore results. . User interface developed GUI postprocessing for both onshore and oﬀshore results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 6.2 6.3 6.4 6.5
35 36 36 40 50 51 51
52
Surface lines heat loss calculation with six diﬀerent insulation materials. 58 Surface Heat Loss calculation without insulation. . . . . . . . . . . . Heat loss from sea level to sea ﬂoor with six diﬀerent insulations. . . 58 59 59
Heat loss from sea level to sea ﬂoor without insulation. . . . . . . . . Heat loss calculation using diﬀerent insulation materials based on Example 10.2 from Prats[41]. . . . . . . . . . . . . . . . . . . . . . . . .
62
6.6
Heat loss calculation without using insulation materials based on Example 10.2 from Prats [41]. . . . . . . . . . . . . . . . . . . . . . . . . 62
6.7
Comparison of steam temperature with ﬁeld data 1 and twophase correlations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.8
Comparison of steam pressure with ﬁeld data 1 and twophase correlations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.9
Calculated steam quality with diﬀerent twophase correlations based on ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.10 Calculated heat loss calculation with insulated tubing based on ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii 66
. . . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 Comparison of our model with Fontanilla’s model steam pressure with ﬁeld data 2. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . 6. . . 6. . . . . . . . . . . . 6. . . 1 year. . . . .16 Comparison of our model with Fontanilla’s model steam pressure with ﬁeld data 1. . . 6. . . . . . . . 6. . . . . . . . . . .22 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . .18 Comparison of our model with Fontanilla’s model steam temperature with ﬁeld data 2. 6. . . .17 Comparison of our model with Fontanilla’s model steam quality with ﬁeld data 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 Comparison of our model with Fontanilla’s model steam quality with ﬁeld data 2. . . . . .14 Calculated heat loss calculation with insulated tubing based on ﬁeld data 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 year. . . . . xix 75 75 72 72 71 70 70 69 69 68 68 67 . . . . . .21 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . . . . . . . . . . . . . . .11 Comparison of steam temperature with ﬁeld data 2 and twophase correlations. . . . . . . . . . . . . .15 Comparison of our model with Fontanilla’s model steam temperature with ﬁeld data 1. . . . . . . . . . . . . . . . . . 6. . . . . .12 Comparison of steam pressure with ﬁeld data 2 and twophase correlations. . . . . . . . . . . . . . . . . . . . 6. . . . . . 6. . . . . . . . .13 Calculated steam quality with diﬀerent twophase correlations based on ﬁeld data 2.6. . . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6. . 1 year. 1 year. . . . . . . .27 Steam quality distribution. . . . . . . . . 1 year.25 Steam temperature distribution . . . . . . . . . . . . . . . . . . . . . . 6. . . . . .28 Heat loss distribution. . . . . . .24 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . .31 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . 6. . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel. . . . . . . . . . . . . . . . . . . . . . . .6. . . 6.23 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . 1 year. . . . . . . . . . . 6. . . . . . . . . . . . . . 6. . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation.30 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel. . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . 6. . . . . . . . . . . . . 1 year. . . . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . . 1 year. . .33 Steam temperature distribution. 79 6. . . . . . . 1 year. . . .32 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . xx 81 80 80 79 78 78 77 77 76 76 . . . .29 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . 6. 1 year. . . . . .26 Steam pressure distribution. . . . . . . . . 1 year. . 1 year. . Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel. . . . . . . . . . . . . . . . 1 year. . . . . Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel. 6. . . . . . .
. . . . . . . . . . . . . . . . 1 year. 6. . . . . . . . . . . 1 year. . . . 6. . . . . . . . . 6. . . .43 With changing N2 molar percentage. . 6. 6. . . . . . . . . . 1 year. .41 With changing injection temperature.42 With changing injection depth. . . . . . . . . . . . . 1 year. . . . .44 With changing injection rate. .39 With changing injection rate. . . . . . .38 With changing N2 molar percentage. . xxi 93 91 90 89 88 87 86 85 84 83 82 82 81 . . . 1 year. . . . . . . . . . . . . . . . . . . . . . . 1 year. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . 6.35 Steam quality distribution. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . 6. . . . . . . . .6. 6. . . . . . . . . . . . . . . . . 1 year. . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. 6. . . . .45 With changing steam quality molar percentage . . . . . . . . 1 year. . . . . 6. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . . . . . . 1 year. . . . . 1 year. . . .46 With changing injection temperature. . . 6. 1 year. . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. 6. . . 1 year. . . . . . . . .37 Pressure drop distribution and formation pressure (green dots). . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . . . . . . . . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . 1 year.34 Steam pressure distribution. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . . . . . . . .40 With changing steam quality molar percentage . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel for onshore. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. .36 Heat loss distribution. . . . . . . . .
. . . . 6. . . . . . 1 year. . . .51 With changing injection temperature. . 111 A. . . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . . . . . .54 With changing injection rate. . . .52 With changing injection depth. . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . 6. . . 6. . . 100 6. . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. .49 With changing injection rate. . . . . . . . . . . 1 year. . 126 A. . . . . 1 year. . . . . . . . . 129 A. . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . . . . . .53 With changing N2 molar percentage. . . . . 6. . . 1 year. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . . . . . .47 With changing injection depth. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. . . .50 With changing steam quality molar percentage . . . . . . . . . . . . .1 Schematic view of tubing element in our calculation. . . . . . . .6. . 1 year. . . . 1 year. . . . .48 With changing N2 molar percentage. . . 100 6. . . . . . . . 101 A. . . . . . 1 year. .3 Viscosity of the annular ﬂuid with respect to Temperature. . . . . . . . 1 year. .55 With changing steam quality molar percentage . . . . . 1 year. 6. . . . . . . 1 year. . 6. . . . . 1 year. . 6. . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . . . . . . . . 130 xxii 99 99 98 97 97 96 96 94 . . . . .56 With changing injection temperature. .2 Time conduction function (retrieved from[19]). . .57 With changing injection depth. . . .4 Thermal conductivity of the annular ﬂuid with respect to Temperature. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.
159 C. . . .8 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. 1 year. . . . 1 year.2 Schematic representation of resistance to heat transfer with temperature proﬁle. . . . . . . . . . . . . .5 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . 137 B. . . . . 1 year. . . . . 1 year. . .6 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . .B. . . . . . . . . . . . . . . .1 Steam temperature distribution for diﬀerent injection temperature vs depth (ft) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass.4 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. 1 year. . . . . . . . . . . 158 C. 145 C. . . . . . . . . . . . . 157 C. 1 year. . . . .3 Steam quality distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. . . . . . . . . . . . . . . . . . . . . . . .2 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 158 C. . . . . . . . . . . . 156 C. . . . .1 Schematic representation of resistance to heat transfer with/without temperature proﬁle. .7 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . . . . . .159 C. . . . . . . . . . . . . . . . . . . . . . . . . . 160 xxiii . . 157 C. . . 132 B. . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. . . . . . . . 1 year. . . 1 year. . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . .3 Schematic representation of resistance to heat transfer sea ﬂoor to reservoir. . . .
. . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . . . . . . . . . . . . . 1 year. . 1 year. .17 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . .12 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . . . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . 163 C. .14 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation.10 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. . 163 C. . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. . . 164 C. . . 161 C. . . . . . . .11 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . .C. . 164 C. .13 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . .15 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . . . . . 161 C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 year. . . . . . . . . .16 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . . . . . . . . . . . . 165 xxiv . . . 1 year. 1 year. . . . . . 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . . . . . . . . . . . . . . . . . . 1 year. . 1 year. . . . . .9 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . . . 160 C. . . . . . . . . . . . . 162 C. . . . .
. . . . . . . . .19 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . 166 C. . . . 167 C. . . .22 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . 168 C. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. 167 C. . 1 year.18 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . 1 year. . . . . . . . . . . . . . . . . 1 year.170 xxv . .26 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . 1 year. . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . . . . 168 C. . . . . . . . . . . . . . . .C. . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. . 1 year.27 Steam quality distribution for diﬀerent injection temperature vs depth (ft). 1 year. . 1 year. .24 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . 165 C. . . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . . Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. . . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . . . .25 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . 1 year.23 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . .20 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . . . . . . . . . . 169 C. . . . . . . . . . . . . . . . . . . 1 year. . . . .169 C. . . . . . . . . . . . .166 C. .21 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . . . . . .
1 year. . . . . . . . . . . . . . .31 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . . . . .35 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. 174 C. . . . . 1 year. . . . . 1 year. . . . . . . . 173 C. . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. . . . . . . 172 C. . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. . . . . . .28 Heat loss distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . . . . . . . . . . . . . . . . . . . . .33 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . .32 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . 173 C. . . . . . . . . 1 year. . . . . . . . . . . . 1 year. . 1 year. . . . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. 171 C. . . . .C. . . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. . 1 year. . . . . . . 1 year. . .29 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. .34 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . . .30 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . . . 171 C. . . . . .36 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . . . 170 C. . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. . . . . 1 year. . . . . . . . . . . . . . . . 174 xxvi . . . . . . 172 C. . .
. . . . . . . . . .176 xxvii . . . . . .38 Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 175 C. . . . . . . .C. . . . 1 year. . . .37 Steam temperature distribution for diﬀerent injection temperature vs depth (ft). . . . . . .39 Steam quality distribution for diﬀerent injection temperature vs depth (ft). . . . . . . . . . 176 C. . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . . . . . . . 1 year. . . 175 C. 1 year. . . . . . 1 year. . . . . . . . . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . . . . . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . . .40 Heat loss distribution for diﬀerent injection temperature vs depth (ft). . . . . Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. . .
xxviii .
The displacement of ﬂuids by steam is self stabilizing movement of the crude oil to the production well. Many investigators have studied heat transfer to the surrounding formation while 1 . viscosity decreases. are easily the most successful enhanced oil recovery processes. Prats[41] indicated that the eﬀect of steam injection on recovery is signiﬁcantly greater as compared to hotwater injection. It was the ”Emeraude Vapeur” pilot test that had great technical success [9]. but only one example exists in the literature for oﬀshore ﬁelds. Crudeoil viscosity is inversely proportional to temperature. provide drive energy and thereby improve the displacement eﬃciency of injected ﬂuid [44] . Lake[30] emphasized that thermal methods especially steam injection and steam soak. Steam injection is the thermal method that add heat to the reservoir to expand the oil inplace. Many applications of steam injection have been done with success in the onshore environment. It is because steam carries more enthalpy per unit mass. Less viscous oil results in greater mobility. reduce oil viscosity. When the temperature increases.Chapter 1 Introduction Increasing oil prices have helped raise investment in EOR applications during the last two decades. Steam injection is applied to viscous oil reservoirs in order to reduce oil viscosity and increase production.
temperature losses. and single phase. He neglected kinetic energy. INTRODUCTION hot ﬂuid injection travels downward along the wellbore. For the onshore cases. noncompressible. Heat loss and pressure drop calculation for steam injection in oﬀshore environments are not reported in the literature. Several twophase correlations in the literature are used. Willhite [49] proposed a well known overall heat transfer coeﬃcient calculation that has been widely used in the oil industry since. Ramey assumed that ﬂow is steady state. Earlougher [17] considered steam and casing conditions with respect to depth. Several years later Pacheco and Farouq Ali [38] presented a comprehensive mathematical model of steam injection without taking into account slip and the multiphase ﬂow regime concept. For all calculations. He solved energy and momentum balance equations analytically to get approximate pressure and temperature distributions. we start with a base case following . pressure losses. frictional loss. pressure drops and quality changes along the wellbore are calculated by coupling nonlinear equations and solving them simultaneously. and steam quality changes along the wellbore during steam injection in both onshore and oﬀshore environments. He used the Hagedorn and Brown model [22] for twophase ﬂow calculation and assumed that there is no slippage between the phases for steam injection. He took into account the slip concept and ﬂow regime of the ﬂow and concluded that considering the slip and the ﬂow regime is important for calculating pressure drop and steam quality during the downward steam injection. Farouq Ali [18] solved this issue and proposed a comprehensive wellbore steam ﬂow model. He actually assumed that the gas and liquid phase ﬂow at the same speed. In 1967. and assumed that the overall heat transfer coeﬃcient does not change with depth. In 1969. Ramey [42] was the pioneer and his model is used by most researchers as a starting point. The objective of this study is to calculate the heat losses.2 CHAPTER 1. Early in the 1980s. Oﬀshore cases take into account the thermophysical properties of sea water to get the correct radiation and convection heat transfer coeﬃcients.
Figure 1. This is the ﬁrst time that this kind of calculation is described in the open literature. however. such as N2 . In order to calculate heat losses from the oﬀshore environments you must consider surface lines.1 shows the schematic view of the our calculation for both onshore and oﬀshore environments. insulating small amount of length gives much eﬃciency on steam quality. pressure drops and steam quality change along the wellbore during injection of steam with an additive of noncondensable gas. then extend our work by applying several more updated twophase correlations. sea level to sea ﬂoor and sea ﬂoor to the reservoir. Surface lines does not contribute heat losses with comparing sea level and sea ﬂoor to reservoir. The results obtained using diﬀerent twophase ﬂow correlations are compared. we conduct calculation of heat transfer.3 similar procedures as shown by Fontanilla[20]. . As another step forward. A novel approach is proposed to solve this problem. Sensitivity analysis using diﬀerent insulation materials are also conducted to investigate the eﬀect of insulation materials on downhole steam properties.
Then. the wellbore heat transmission concept is reviewed starting from Ramey’s [42] classical paper. INTRODUCTION Figure 1. we discuss model formulation that we used in our calculations. . One case had great success in terms of producing more oil resulting from steam injection. we review a summary of three basic heat transfer mechanisms and their combinations that are the backbone of our study. we review two cases of heat ﬂow from wells for oﬀshore ﬁelds.4 CHAPTER 1. 1.1: Schematic view of the objective of our calculations. In addition. (retrieved from [2]). In Chapter 3. the other case used vacuum insulating tubing (VIT) to reduce heat losses and failed.1 Thesis Outline In Chapter 2.
we also provide a robust solution that takes into account thermo physical properties of seawater in order to calculate important parameters [36] for heatloss calculations. We augment Fontanilla’s approach as well. For other correlations. it causes a decrease in steam pressure and so steam temperature. references are provided. With this program a user can study the role of several parameters and see the eﬀects of those parameter on the system. and sea ﬂoor to reservoir based on both Prats’s [41] and Willhite’s [49] algorithm. THESIS OUTLINE 5 Starting with heat transfer calculation in three parts: surface lines. Once N2 is injected into the system. The N2 gives additional pressure into our steam and helps to reduce heat loss. Continue with steam properties calculation. we present and discuss most of the twophase correlations that can be applied for downward steam injection operations. Chapter 6 discusses the results obtained in preceding chapters and compares them in terms of using diﬀerent insulation materials and steam properties. In order to achieve this purpose.1. In Chapter 4. we calculated steam and noncondensible gas properties for insulated and noninsulated tubing in both onshore and oﬀshore environments. the problem of steam with noncondensable gas (N2 ) is considered.1. Besides the analogy for oﬀshore. This program allows a user to choose several correlations to be applied such as insulated or uninsulated tubing. Several of correlations are explained in detail. sea. Chapter 5. We validated our program with ﬁeld data from the literature [11] and obtained good agreement with ﬁeld data and also with Fontanilla’s approach [19]. . A summary of our ﬁndings and suggestions for future work are presented in Chapter 7. is the one of the exciting parts of this thesis because we developed a Graphical User Interface (GUI) for our calculations and gave the basics of the program input and output that is used in our calculations.
Figure 2.1 a. These very heterogeneous reservoirs are signiﬁcantly depleted and oil is viscous (0. The ﬁrst is Emeraude Vapeur[9] and subsequently the Marlin failure and redesign [16. on the West African coast. and adapted pumping units on one platform because of reservoir 6 . The depth of reservoirs is shallow (200500 m) and they consist of silt layers alternating with thin fractured limestone beds. 48] are discussed.Chapter 2 Literature Review We review two examples from the literature for oﬀshore cases of wellbore heat losses. a tilted rig. Water depth is 65 m. 2. The adverse environmental conditions required original solutions: tilted conductor pipes. A steam drive pilot test was decided in order to estimate a recovery rate and an oilsteam ratio on two independent reservoirs in 1980.1 Emeraude Vapeur : A Steam Pilot in an Oﬀshore Environment The Emeraude ﬁeld is located oﬀshore Zaire (Congo). 43. This chapter continues with heat transfer mechanism and with heat transfer discussions.1 Pa s (100 cp)) at reservoir conditions.
(b) Emeraude ﬁve spot. were obtained from steaminjection tests under laboratory conditions. assuming that imbibition would be active. with water breaking through to producers almost immediately. Water injection was implemented in 1972 in a ﬁvespot pattern. several additional platforms would be needed. The .1: Emeraude ﬁeld location and ﬁve spot[9]. After 14 years of production (19721986). and the reservoir was severely pressure depleted.1. of this large amount of oil associated with this poor recovery rate. Figure 2. about 3% of OOIP. however. EMERAUDE VAPEUR : A STEAM PILOT IN AN OFFSHORE ENVIRONMENT7 shallowness and steam production equipment on a second platform because of the distance to the shore. Various EOR methods were considered to meet the challenge (a) Emeraude ﬁeld location. Promising results. only 170 million barrels had been recovered. The Emeraude ﬁeld is estimated to contain 1 billion barrels of viscous original oil in place (OOIP). Insitu combustion tests under laboratory conditions showed that most of the oil would be burned in the fracture network. and the ﬁnal recovery would still be only 510% OOIP [9]. The results were disappointing. To produce the remaining reserves by primary recovery in 1520 years.2.
facilities for steam production from seawater were located on a platform because of the long distance to shore. and OSR (oil steam ratio). steam can be injected at a suﬃcient rate in reservoirs R1 and R2. One of the ﬁrst developments was obtained interms of getting suﬃcient well spacing by using a tilt rig for drilling. drilling through and cementing depleted and fractured zones. steam injectivity.[15] shows that both steam and hotwater drives may improve oil mobility by reducing viscosity and also may reduce residual oil at high temperatures. located on adjacent platforms were noted. pumping tilted wells. steam breakthrough time. some reservoir engineering conclusions were obtained. A third was that pumping units were adapted to tilted wells and electronically regulated. In addition to these technological solutions. A ﬁnal conclusion was that steam improves the oil production rate in heterogeneous reservoirs. A steamﬂood in Figure 2. The Emeraude steam drive pilot provided original solutions to recover a larger amount of OOIP than could be recovered by primary production despite diﬃcult conditions. Lastly. • prove the technological feasibility of such a project in adverse conditions (drilling with a tilted rig. LITERATURE REVIEW experimental work of Willman et al. recovery rate. Signiﬁcant response by Well EMV07 in reservoir R2 (oil rate increased fourfold) and other wells. A second success of development was that depleted and fractured zones were adequately drilled and completed. For instance.1 b shows the ﬁve spot steam injection implemented to.8 CHAPTER 2. and producing steam from seawater) and • evaluate reservoir responses to steam injection. . Technological problems were solved during the pilot design.
and hydrogen released from the inner tubular at elevated temperatures [6]. VIT provides a solution for heat loss in steam injection in Arctic and oﬀshore environments. It was the ﬁrst alert of a major tubing failure.2. Right after production began. Ellis et al. three papers are reviewed related to this subject and including basic information about Vacuum Insulated Tubing (VIT). The Marlin ﬁeld is located in the Gulf of Mexico. carbon monoxide. and was originally intended to be produced from a tensionlog platform by means of ﬁve predrilled drytree penetrations. The annulus is usually ﬁlled with a better material for absorbing hydrogen formed by corrosion of the outer tubular and gases such as nitrogen. MARLIN FAILURE ANALYSIS AND REDESIGN 9 2. They try to reduce them by making analysis and physical evidence. VIT is a tubular apparatus conveying steam or other hot ﬂuids (> 400o F ) to the formation through an inner tubular that is surrounded by an outer tubular [6]. [16] describe the Marlin failure and give several possible failure modes. • hydrate formation and dissolution. Gosch et al. Viosca Knoll Blocks 871/915. Analysis of the failure came up with several critical issues including • excessive helical buckling of the production tubing.2. a minor tubing leak occurred. and .2 Marlin Failure Analysis and Redesign In this section. Bradford et al. The annular space between the inner and the outer tubular is under vacuum. and after that casing pressure jumped to shut in tubing pressure. • trapped annulus pressure leading to casing collapse. [43] apply the failure analysis from the ﬁrst paper to the remaining Marlin wells and focuses on the VIT redesign process. [48] addresses focusing on the value of combined VIT and ﬁber/software monitoring system as a means of both controlling and observing well thermal behavior.
a combination of thermal coatings and insulated inserts provides adequate additional insulation at the couplings. Using VIT. VIT design itself has lots of design considerations and challenges. Another important conclusion was natural convection can signiﬁcantly aﬀect the ability of VIT to isolate tubing temperatures from the annulus. They still do not know whether it was because of the production tieback collapsed alone or as a result of collapse of the intermediate casing is unknown [16].10 CHAPTER 2. and environment (HSE). Second. Although Marlin wells predrilled up to the completion stage limited mitigation options. safety. The overall thermal properties of the unaltered VIT were not adequate for Marlin requirements due to the fact that heat loss at coupling dominates the performance of a VIT joint. several materials were tested in the annulus to reduce heat loss and N2 is the most eﬀective barrier to heat loss. they had signiﬁcant thermal isolation due to the low conductivity or low convecting annular ﬂuids [43]. Additionally. well design concepts were developed and screened using agreedupon riskacceptance criteria for health. They chose VIT based on the economic analysis and risk proﬁle associated with each option. it . a well is shut in automatically. They sum up their work in several with giving results. Last. They tried to understand whether the VIT was a good choice or not based on the both experimental and numerical results [48]. Design of VIT introduces a number of considerations not present in a design using conventional tubing. This realtime monitoring gave them better control of the well such as when a low safety factor is calculated. they have tested VIT performance with developed software to see the production annulus temperature proﬁle in real time. In the third paper of the Marlin failure redesign [48]. they concluded that deformation of Well A2’s tubing was the result of collapse of production tieback. They did several experimental studies on VIT and found out three important facts. LITERATURE REVIEW • improper tubulars and wellhead movement After the ﬁshing operation and ultrasonic caliper.
3. Both the thermal performance of the VIT and its mechanical integrity require special consideration. convection.1) where k is the thermal conductivity of the material (Fig.3 Heat Transmission Mechanisms and Discussion from Authors 2. reduces to the diﬀerential form that is . less energetic ones as a result of interaction between particles. and radiation and a combination of two or more in our calculations. T1 − T2 ∆T ˙ Qcond = kA = −kA ∆x ∆x (2.1.2: Schematic view of conduction ing case ∆x → 0 the equation above (after [8]). HEAT TRANSMISSION MECHANISMS AND DISCUSSION FROM AUTHORS11 was remarked that regional heating can add a temperature increase to outer annuli that is not anticipated in a single well analysis.1 Heat Transfer by Conduction Heat conduction also called diﬀusion is the transfer of energy from the more energetic particles of a substance to the adjacent. 2. In the limitFigure 2.2). Including conduction. These results lead the authors to suggest that each well has to studied and treated individually.3. 2. 2.2.3.1 Heat Transmission Mechanisms In this section we review the heat transfer mechanisms during steam injection operations.
35]. 2.12 CHAPTER 2. In contrast.2 Heat Transfer by Convection (2. The radiation that can be emitted from a surface at an absolute temperature Ts (in K or R) is given by Figure 2. Fourier and becomes [8. dT ˙ Qcond = −kA dx 2.max = σAs Ts4 (2.3. pump.3 Heat Transfer by Radiation Radiation is the energy emitted by matter in the form of electromagnetic waves as a result of changes in the electronic conﬁgurations of the atoms or molecules. 35] by convection (retrieved from [1]).4: Representation the Stefan − Boltzman law as [8.3). or the wind.3) 2. Convection is called forced convection if the ﬂuid is forced to ﬂow over the surface by external means such as fan. ˙ Qemit. The rate of convection heat transfer is expressed by Newton s Figure 2. LITERATURE REVIEW called Fourier s law of heat conduction after J.2) One mode of convection heat transfer is between a solid surface and the adjacent liquid or gas that is in motion.1.4) .3: Heat transfer from a hot surface to air law of cooling as [8. 35] of heat transfer by radiation(after [8]).3. convection is called natural (or free) convection if the ﬂuid motion caused by buoyancy forces that are induced by density diﬀerences due to the variation of temperature in the ﬂuid (Fig.1. ˙ conv = −hAs (Ts − T∞ ) Q (2.
He assumed that ﬂuid is noncompressible and ﬂow is single phase with constant thermal and physical properties along the wellbore. Holst and Flock [25] added the friction loss and kinetic energy eﬀects on Ramey’s [42] and Satter’s [45] models. however. One of the wellknown papers for wellbore heat transmission is by Ramey[42].3. Satter [45] presented a method that improved Ramey’s [42] model by making the overall heat transfer coeﬃcient dependent on depthstep method for calculating heat loss and steam quality for saturated steam as a function of depth. HEAT TRANSMISSION MECHANISMS AND DISCUSSION FROM AUTHORS13 2. he assumed. Leutwyler [31] gave a comprehensive treatment of casing temperature behavior. He did not take into account frictional pressure loss and kinetic energy eﬀect in his calculation. One year after Satter’s [45] paper. Since Satter assumed that there is no change in pressure with depth. In his paper.2 Heat Transmission Discussion from Authors In the literature many investigators have worked on the thermodynamic properties of the hot ﬂuid movement through wells in onshore ﬁelds for both production and injection. Squier et al.3. Hans and Huitt [26] also developed a graphical solution for wet steam . in order to calculate the heat loss and quality distribution versus depth for saturated steam injection operations. using a complete analytical method. In 1966. They neglected. the static pressure change. Most of the publications follow upon his approach.2. [47] solved diﬀerential equations describing ﬂuid temperature along the wellbore. in eﬀect that the temperature of the injected saturated steam remains constant. He provided an approximate analytical solution for wellbore heat transmission. Ramey [42] made several assumptions. They used hot water as injection ﬂuid. He considered that heat ﬂows radially away the wellbore and the overall heat treansfer coeﬃcient is independent of depth. and that only the quality varies with depth.
and casing temperature. Their approach was assuming nonhomogeneous formations as layered formation with diﬀerent physical properties. He used several correlations and stated that importance of applying twophase ﬂow concept and ﬂow regime. Wu and Pruess [50] presented a new analytical for wellbore heat transmission without Ramey’s assumptions. All the development both technological and understanding of the problem have been done so far helped to understand heat transfer mechanism and solve the problem with taking into account considerations from the authors. Willhite [49] proposed his well known method for estimation of overall heat transfer coeﬃcient that is applied in our calculation as well. Two of the pioneers in the prediction of heat loss and pressure drop in the wellbore were Pacheco and Farouq Ali [38]. and later on Farouq Ali [18] solved this problem by taking into account slip between the ﬂuids and the ﬂow regime. LITERATURE REVIEW injection operations. they calculate wellbore heat loss. . which is not valid. One year later. They formulated a mathematical model that consisted of two coupled nonlinear diﬀerential equations that were solved iteratively in terms of pressure and quality of steam. In their model. steam condensation rate. They assumed single phase ﬂow.14 CHAPTER 2.
The tubing hangs symmetrically inside the casing. 3. 15 . This can be seen in the derivation of the equation to evaluate the overall heat transfer coeﬃcient Uto . Heat transfer in and around the wellbore takes place under pseudo steady state conditions. These assumptions follow: 1.Chapter 3 Model Formulation This chapter presents the methodology for heatloss calculations and pressure drop estimation in injection tubing.1 Heat Loss Calculations We have adapted Fontanilla’s [19] assumptions for the solution of oﬀshore and onshore injectors. This would mean that the heat rate through the wellbore components is the same at any given time. The heat transfer into the earth occurs under unsteadystate condition. 3. 2. Conduction is pseudo steady state when the change in temperature with respect to time at any given point in the wellbore remains constant. The heat diﬀusivity and the conductivity of the formation is constant.
The procedures for estimating heat losses may appear laborious. it loses energy to the surrounding formation. whether surface lines or wells.1: Thermal Conductivity of the materials Insulation Materials Black Aerogel White Aerogel Fiberglass Carbon Fiber Thermolastic Insulation Calcium Silicate W/(m*K) 0.028 0. In Table 3.014 0.1.0400 .0237 0. This heat loss may result in condensation with consequent reduction in steam quality and enthalpy. The calculation procedures of both Prats [41] and Willhite [49] are discussed here. For comprehensive calculation procedure we refer to read Appendix A. One of the classic papers about overall heat transfer coeﬃcient was published by Willhite [49]. Heat losses through pipes. we have provided thermal conductivity of the diﬀerent insulation material based on the values from the thesis of Marques[34].041 0.0162 0. Here in our study.0208 0. He presented his widely used method that is incorporated in most simulators for hot ﬂuid injection and hot water. we will give the equations for both surface lines and wellbores based on the methods discussed. usually are estimated at steadystate conditions in oilﬁeld operations.16 CHAPTER 3. MODEL FORMULATION As steam travels downward in the wellbore. In addition.0069 0.036 0. Table 3.012 0.069 (BT U/(f t − hr −o F)) 0.0081 0. For oﬀshore heat losses calculations an analogy between electrical circuits and heat resistance is made. we provide the equations for insulated and noninsulated cases and also for oﬀshore heat loss calculations.
In Figure 3. (b) Surface pipe without insulation.1. HEAT LOSS CALCULATIONS 17 3. and . is Tb − TA ˙ Qls = Rh where Rh is represented as Rh = 1 . Qls .1 a. ˙ The basic equation used to calculate heat losses per unit length of pipe.3. we have a representative resistance to heat transfer and a temperature proﬁle. That savings can be signiﬁcant will be demonstrated by means of steam injection examples. it is generally worthwhile to use insulation to reduce heat losses to save both money and fuel.1: Schematic representation of resistance to heat transfer with or without temperature proﬁle.1.1.1. Figure 3. Prats [41] stated that even though heat losses from surface lines in hot ﬂuid injection operations may be a small fraction of the total heat injected.1 Heat Loss from Surface Lines With/without Insulation (a) Surface pipe with insulation. 2πrU (3.1 3.1) U is the overall coeﬃcient of heat transfer.
radiation is usually insigniﬁcant and is not included in Eq. and λp and λins are the thermal conductivities of the pipe and insulation. and Qls is the rate of heat loss per unit length of pipe in BT U/(f t − hr). the speciﬁc thermal resistance of heat loss is given as Rh = 1 1 1 1 ro 1 1 rins 1 + + ln + + ln + 2π hf ri hpi ri λp ri hpo ro λins ro hf c rins (3. The physical signiﬁcance of each of the six terms in the right side of Eq.1. hf c is the coeﬃcient of heat transfer due to forced convection (air currents) at the outer surface of the insulation.3.2 is illustrated in Figure 3. hpo is the coeﬃcient of heat transfer across the contact between pipe and insulation. rins is the external radius of the insulation. For a pipe covered with insulation. Because the temperature on the surface of most insulated lines is low. Tb is the bulk temperature of the ﬂuid in the pipe in degrees Fahrenheit.2) Here hf is the ﬁlm coeﬃcient of heat transfer between the ﬂuid inside the pipe and the pipe wall. Rh is the speciﬁc thermal resistance (thermal resistance per unit length of pipe) and is given in units of (BT U/(f t − hr −o F))−1 .ro is the outer radius of the pipe and essentially the inner radius of the insulation. because the transient phase is often of short duration (of the order of less than a day). Coeﬃcients of heat transfer are expressed in (BT U/(sqf t − hr −o F)) . Each of the six terms is proportional to a thermal resistance in the system aﬀecting heat losses. and thermal conductivities in (BT U/(f t − hr −o F)). radii in feet.18 CHAPTER 3. TA ˙ is the ambient temperature of the atmosphere in degrees Fahrenheit. Adjacent to the inner surface of the pipe is a lowvelocity .2 above. Rates of heat loss during transient periods can be several times greater than at steady state. hpi is the coeﬃcient of heat transfer across any deposits of scale or dirt at the inside wall of the pipe. Transient eﬀects generally are neglected in calculations of heat losses from thermal lines. ri is the inner radius of the pipe. MODEL FORMULATION r is an arbitrary radius that usually coincides with the radius of one of the surfaces for which the heat loss is being determined.3. Here.
leads to the coeﬃcient of heat transfer hf c . which aﬀects heat losses to the atmosphere by forced convection.2 3. It should be pointed out that adding more insulation does not necessarily reduce the rate of heat losses further.2. . Note that the resistance to heat ﬂow across this ﬁlm decreases as the value of the coeﬃcient of heat transfer increases. 3. Figure 3. (b) Sea Part without insulation.1.2: Schematic representation of resistance to heat transfer with temperature proﬁle.1. Scale or dirt deposits at the inside (2) and outside (4) pipe walls lead to coeﬃcients of heat transfer hpi and hpo . Heat transfer through the pipe wall (3) and the insulation (5) is by conduction.1.3.1 Heat Loss from Sea Level to Sea Floor With/without Insulation (a) Sea Part with insulation. Because of its low velocity. this ﬁlm has heat transfer characteristics diﬀerent from those of the ﬂowing bulk ﬂuid and accounts for the introduction of the ﬁlm coeﬃcient of heat transfer hf . HEAT LOSS CALCULATIONS 19 ﬂuid ﬁlm (1). respectively. A low velocity ﬂuid ﬁlm at the exterior surface of the insulation (6).
2.4 as reference: 1. 4. heat transfer is steady because the speciﬁed thermal conditions at the boundaries do not change with time. heat transfer is one dimensional due to thermal symmetry about the midpoint. thermal conductivity is constant. sea temperature does not change along the wellbore Rtotal = Rconv. density. 3.1 + Rcycl.1 + Rcycl.3) (3.2 Rtotal = 1 1 r2 1 r3 1 r4 1 + ln + ln + ln + h1 A1 2πLk1 r1 2πLk2 r2 2πLk3 r3 h2 A4 (3. we have made several assumptions taking Figure 3.2 + Rcycl. MODEL FORMULATION For the sea level heat loss calculation.20 CHAPTER 3.3 + Rconv. thermal conductivity. speciﬁc heat at the given salinity and temperature of the sea.4) In order to calculate heat losses from oﬀshore wells we have to ﬁnd Tins and Uto and also sea water parameters such as thermal diﬀusivity. .
1. Figure 3.4: Schematic representation of the wellbore.3: Sea water properties change with temperature and salinity[36]. HEAT LOSS CALCULATIONS 21 Figure 3.3. .
thus. For the insulated tubing held concentrically within the casing shown in this .22 CHAPTER 3. A representation of the typical elements oﬀering resistance to heat losses from the wellbore is given in Fig 3. where in this case the ambient temperature is the geothermal temperature and.1.1. Figure 3. (b) Sea Floor to Reservoir without insulation. the speciﬁc thermal resistance is time dependent. is discussed later in more detail. Heat losses from wells never reach a steady state.3 3. It is a measure of how fast the earth conducts heat away from the well. This function of time.1. 3. MODEL FORMULATION 3. of course. reﬂecting the variable eﬀective thermal resistance of the earth.1 Heat Loss from Sea Floor to Reservoir With/without Insulation (a) Sea Floor to Reservoir with insulation. They attain. a function of depth.5. Heat losses from the well to the earth are characterized by Eq. In this case. a quasisteady state in which the rate of heat loss is a monotonically decreasing function of time.3. as pointed out by Ramey [42] and Willhite [49].5: Schematic representation of resistance to heat transfer sea ﬂoor to reservoir.
f (tD ) is represented in terms of dimensionless time: tD = αE t 2 rw (3. hς. rEa is the radius of the altered zone in the earth near the wellbore. Coeﬃcients of the heat transfer are expressed in (BT U/(sqf t − hr −o F)). rw is the wellbore radius. and f (tD ) is the time function that reﬂects the thermal resistance of the earth. λcem is the thermal conductivity of the cement. λEa and λE are the thermal conductivities of the altered and unaltered earth. and the dimensionless time is discussed later. and the variable resistance of the earth. rci and rco are the inner and outer casing radii.5) The ﬁrst ﬁve terms have been discussed in the preceding section for heat loss from surface lines. here only the Ramey[42] and Willhite[49] representation is going to be discussed. In Eq. The last ﬁve terms represent. in order of appearance.an . The function f (tD ) has been discussed by a number of authors. The function f (tD ) and the radiationconvection coeﬃcient of heat transfer in the annulus. the resistance to radiation and convection in the annulus.an is the radiation and convection coeﬃcient of heat transfer for the annulus. and thermal conductivities in (BT U/(f t−hr−o F)). radii in feet. the resistance of an altered zone (resulting from drying due to high temperatures) in the earth. the resistance of the cement. The function f (tD ) is dimensionless.1. Diﬀerent well designs lead to diﬀerent expressions for determining the overall thermal resistance Rh . the heat resistance elements are combined to obtain the overall coeﬃcient of heat loss: Rh = 1 1 1 ro 1 rins 1 1 1 + + ln ri + + ln + 2π hf ri hpi ri λp hpo ro λins ro hς.3.6) . the resistance of casing.an rins 1 rco 1 rw rEa f (tD ) 1 + ln + ln rco + ln + λp rci λcem λEa rw λE (3. are the only additional terms requiring discussion. hς. HEAT LOSS CALCULATIONS 23 ﬁgure.
2.7) here αE is the thermal diﬀusivity of the earth in square feet per hour. . For values of tD ≤ 100 and is then Ramey [42] gives calculation: f (tD ) 1 ∗ (lntD ) + 0.8) for tD ≤ 100. This Table 3. MODEL FORMULATION if there is an altered zone. and t is the time from start of heating in hours. For values of of tD ≤ 100.403 2 (3.24 CHAPTER 3. tD = αE t 2 rEa (3. Willhite [49] has published Table 3.2 is used to interpolate the value for ﬁnding f (tD ).
240 1.313 0.890 0.806 1.02 0.427 0.580 1.370 1.660 1.0 0.590 1.050 2.619 0.740 10 0.730 2.0 2.0 0.430 0.910 1.323 0.480 1.1 0.900 25 .313 0.984 1.588 0.745 0.330 0.314 0.860 0.890 0.1.750 5.806 1.110 1.5 0.400 2.160 2.870 2.570 2.150 1.000 1.390 2.644 0.880 2.690 2.2: Time Function f (tD ) for the boundary condition model [49].570 2.568 0.802 1. HEAT LOSS CALCULATIONS Table 3.452 0.578 0.870 2.510 2.811 1.370 1.0 10.438 0.880 0.400 1.439 0.0 20.2 0.090 2.240 1.417 0.670 1.770 2.803 1.170 2.990 2.650 1.623 0.423 0.05 0.316 0.616 0.840 2.345 0.570 2.473 0.424 0.560 2.810 1.2 0.020 1.772 0.0 5.698 0.0 100.580 2.860 2.860 2.396 0.580 2.570 1.120 2.970 2.150 2.390 2.250 1.520 1.423 0.538 0.872 1.420 2.958 1.0 50.240 1.440 1. tD 0.080 1.890 0.666 0.020 1.380 1.0 0.590 1.480 2.445 0.842 1.511 0.811 1.572 0.000 2.540 2.617 0.770 2.3.890 0.730 20 0.373 0.01 0.020 1.010 1.560 1.960 2.802 1.840 0.440 2.220 1.170 2.1 0.590 1.433 0.970 2.040 1.810 2.790 1.730 50 0.0 0.050 1.0 100 0.820 1.5 1.010 1.629 0.030 1.160 2.318 0.200 1.660 1.360 1.170 2.
Based on the function from[24]. a fourth order RungeKutta method is used with Matlab ”ode45” function. .26 CHAPTER 3. The appropriate ordinary diﬀerential equations are described in Appendix A. our pressure versus enthalpy diagram looks like Figure 3. To solve this diﬀerential equation in each interval of the well.6. The steam properties such as density of the saturated steam and density of the saturated liquid are calculated directly using IAPWS IF97[24].2 Steam Phase behavior calculations Figure 3. MODEL FORMULATION 3.6: Pressureenthalpy diagram (retrieved from [24]). Steam quality changes with depth.
We provide ﬂow regimes in two phase vertical ﬂow in Figure 3. Drift Flux model.3. Figure 3. We give information about the two phase correlations that are applied in our calculations for vertical downward ﬂow with insulated and uninsulated tubing for both an onshore and oﬀshore environments.7. . and Hasan and Kabir correlations. twophase ﬂow behavior is more complex than for singlephase ﬂow.3 Two Phase Flow Correlations Unlike singlephase ﬂow.7: Gasliquid ﬂowpatterns for vertical pipes (retrieved from [12]). liquid always ﬂows faster than the gas or vapor phase. Shear stresses at the pipe wall are diﬀerent for each phase because of their diﬀerent densities and viscosities. we also addressed ﬂow regimes for vertical ﬂow. The phases tend to separate because of diﬀerences in density.3. Govier and Fogarasi. The main diﬀerence between gas and liquid phase is they do not travel at the same speed in the pipe. TWO PHASE FLOW CORRELATIONS 27 3. Aziz. The twophase ﬂow correlations we used in our calculations are modiﬁed Beggs and Brill. Besides. For downward ﬂow.
.9) where ρs = ρl HL + ρg Hg and the deﬁnition for ρs and the density term is used in the acceleration component. Many correlations have been developed for predicting twophase ﬂowing pressure gradients that diﬀer in the manner used to calculate these three components of the total pressure gradient. happens at the wellhead during steamonly injection. Predicting the ﬂow regimes that occur at a given location in a well is extremely important. dp = dz dp dz + el dp dz + f dp dz (3. but the gas phase predominantly controls the pressure gradient [13. MODEL FORMULATION The general pressure gradient equation is 2 g f ρf vm ρvm dvm dp = ρs sinφ + + dz gc 2gc d gc dz (3.28 CHAPTER 3. 23.8 shows ﬂow regime pattern both injection and production of the ﬂuid. 46]. Droplet ﬂow. The pipe wall is coated with a liquid ﬁlm. In this ﬂow. also known as mist ﬂow. the gas phase is continuous and the bulk of the liquid is entrained as droplets in the gas phase. 14. Some investigators chose to assume that gas and liquid phases travel at the same velocity (no slippage between phases) for evaluating the mixture density and evaluate only a friction factor empirically. Acceleration is sometimes negligible and is usually calculated only for high ﬂow velocities. Figure 3.10) acc The pressure drop caused by elevation change depends on the density of the twophase mixture and is usually calculated using a liquid holdup value. Friction losses require evaluation of a twophase friction factor. Others developed methods for calculating both liquid hold up and friction factor and some chose divide the ﬂow conditions into patterns and developed separate correlations for each ﬂow regime. The empirical correlation or mechanistic model used to predict ﬂow behavior varies with ﬂow pattern.
The wall of the pipe always contacts with the liquid phase [13. During slug ﬂow. 23. Neither phase appears to be continuous. 23. [13. [13. The correlations addressed in this study are discerned from each other by taking into account both the slippage eﬀect and ﬂow patterns. The bubbles move at diﬀerent velocities and except for their density. 14. slip velocity and noslip velocity values. Churn ﬂow is the change of continuous gas phase to continuous liquid phase. the pipe is almost completely ﬁlled with the liquid and free gas phase is present in small bubbles. 46]. Liquid Holdup and Slippage Eﬀect When two or more phases are present in a pipe. both as a thin ﬁlm along the pipe wall and as dispersed droplet in the core. twophase ﬂow [33].3. they tend to ﬂow at diﬀerent . viscosity of both phases. 23. 14. 46]. There is clear distinction between gas bubbles and liquid phase like gas phase trapped into largebubbles.3. superﬁcial velocity of both gas and liquid phases. 46].8: Vertical downward give equations of them. TWO PHASE FLOW CORRELATIONS 29 Annular mist ﬂow also as known as annular droplet ﬂow occurs right after the mist ﬂow and is characterized by the axial continuity of the gas phase in a central core with the liquid ﬂowing downward. Here we are going to Figure 3. 14. have little eﬀect on the pressure gradient. Parameters are calculated in twophase ﬂow requires knowledge of several parameters such as liquid holdup.
that is referred to as gas holdup or gas fraction. That is HL = volume of liquid in a pipe segment volume of pipe segment (3. is deﬁned as the ratio of the volume of liquid in a pipe segment divided by the volume of the pipe segment which would exist if the gas and liquid traveled at the same velocity (noslippage). the insitu volume fractions of each phase (under ﬂowing conditions) diﬀer from the input volume fractions of the pipe.9: Liquid Holdup and Slippage eﬀect representation (retrieved from[4]). That is Hg = 1 − HL (3.30 CHAPTER 3. This causes a ”slip” eﬀect between the phases. As a consequence. Typically the phase that is less dense ﬂows faster than the other. These insitu velocities depend on the density and viscosity of each phase.11) Liquid holdup is a fraction that varies from zero for all gas ﬂow to one for all liquid ﬂow. sometimes called input liquid content. Liquid holdup is deﬁned as the ratio of the volume of a pipe segment occupied by liquid to the volume of the pipe segment.12) Figure 3. MODEL FORMULATION insitu velocities. It can . The remainder of the pipe segment is of course occupied by gas. No − Slip Liquid Holdup Noslip holdup.
3.20) (3. The superﬁcial velocity of a ﬂuid phase is deﬁned as the velocity which that phase would exhibit if it ﬂowed through the total cross section of the pipe alone.15) .17) qg AHg (3. Superﬁcial velocity for the gas phase is vsg = Actual gas velocity is vg = Superﬁcial velocity for liquid phase is vsL = The actual liquid velocity is vs = The twophase mixture velocity is vm = vsL + vsg The slip velocity is vs = vg − vL = vsg vsL − Hg HL (3. TWO PHASE FLOW CORRELATIONS 31 be calculated directly from the known liquid and gas from rates as CL = qL qL + qg qg qL + qg (3.19) qL AHL (3.16) qg A (3.14) Many twophase ﬂow correlations are based on a variable called superﬁcial velocity.3.18) qL A (3.13) Cg = 1 − CL = Velocity (3.
22) vsL vm (3.0111 ∗ 10−7 ∗ T 3 (3. downhill.2934 ∗ 10−7 ∗ T 2. many of the published correlations are applicable for ”vertical ﬂow” only.0661 + 2. The Beggs and Brill multiphase correlation deals with both the friction pressure loss and the hydrostatic pressure diﬀerence.3192 ∗ 10−4 ∗ T 1.21) 3. It was developed using 1” and 11/2” sections of pipe that could be inclined at any angle from the horizontal.23) (3. inclined and vertical ﬂow. 39] correlation.484045 − 3.1106 ∗ 10−2 ∗ T − 7. For saturated water viscosity µw = exp[0.1115 ∗ 10−2 ∗ T 0. namely uphill.0085 + exp[−7. [32]. The Beggs and Brill [3.2085 ∗ 10−5 ∗ T 2 +1.85 For saturated steam viscosity µs = 0.1 Modiﬁed Beggs and Brill Model For multiphase ﬂow. horizontal. is one of the few published correlations capable of handling all these ﬂow directions.95 + 1. Not many correlations apply to the whole spectrum of ﬂow situations that may be encountered in oil and gas operations. 14. while others apply for ”horizontal ﬂow” only.3. First the appropriate ﬂow regime for .9 −2. 7.32 CHAPTER 3. MODEL FORMULATION The noslip holdup is CL = Viscosity Viscosity of the both saturated water and saturated steam is calculated based on the correlation of from Liang et al.
From this the friction pressure loss is calculated using ”input” gasliquid mixture properties.3. The liquid holdup.1 FlowPattern Determination The Beggs and Brill correlation needs to identify the ﬂow pattern at the given ﬂowing conditions in order to calculate the liquid holdup and friction pressure drop. Intermittent or Distributed) is determined. and transition (ﬂow pattern included after a modiﬁcation of the original publication that considers the region between the segregated and intermittent grouped patterns). The boundaries between these groups of ﬂow patterns appear as curves in a loglog plot in the original publication by Beggs and Brill. In order to build the ﬂow map. the insitu density of the gasliquid mixture is then calculated according to the appropriate ﬂow regime. This was later revised so that straight lines could be used instead. CL ). We use this modiﬁed ﬂow pattern map in our calculations. to obtain the hydrostatic pressure diﬀerence. The revised lines that deﬁne the boundaries are deﬁned as follows (where * stands for the modiﬁcation of the original curve to a straight line in a loglog plot) .1. the observed ﬂow patterns were grouped as: segregated (stratiﬁed. For this purpose. 3. and hence. distributed (bubble and mist ﬂow). the Beggs and Brill correlation makes use of a horizontal ﬂow pattern map built based on the Froude number of the mixture (Frm ) and input liquid content (noslip liquid holdup. intermittent (plug and slug ﬂow). TWO PHASE FLOW CORRELATIONS 33 the particular combination of gas and liquid rates (Segregated. A twophase friction factor is calculated based on the ”input” gasliquid ratio and the Moody friction factor table using Colebrook equation.3.3. wavy and annular ﬂow).
Therefore.302 L∗ = 316CL 1 −2. Unless the pipe is actually in the horizontal position. to calculate the liquid holdup.0009252CL 2 −1. 0. the Beggs and Brill correlation is not able to recognize the actual ﬂow pattern under the given conditions. MODEL FORMULATION Figure 3.1CL 3 −6.738 L∗ = 0.25) (3.4516 L∗ = 0.5CL 4 (3. we ﬁrst determine the liquid holdup for the horizontal .24) (3.4684 L∗ = 0.26) (3.27) The identiﬁed ﬂow pattern is the one that would exist if the pipe were horizontal.10: Flow Map for the Beggs and Brill Correlation (retrieved from[4]).34 CHAPTER 3.
4 and L3 < F rm ≤ L∗ or CL ≥ 0. The Froude number is a dimensionless number that relates the inertia with respect to the gravitational forces. IntermittentF low occurs when 0.4 1 and L∗ < F rm ≤ L∗ .11: Segregated Flow Regime (retrieved from[4]). the corresponding ﬂow pattern is identiﬁed when the following inequalities are satisﬁed. For a mixture.01 and F rm < L∗ 2 1 .01 ≤ CL < 0. and this value is then corrected for the angle of interest.3.3. SegregatedF low occurs when CL < 0.01 and F rm < L∗ or CL ≥ 0. it is obtained as F rm = 2 Vm gD (3. Figure 3. 3 4 .28) Once the input liquid content (CL ) and Froude number of the mixture (Frm ) are determined. TWO PHASE FLOW CORRELATIONS 35 ﬂow.
4 and F rm > L∗ .4 and F rm ≥ L∗ or CL ≥ 0. Beggs and Brill divided the liquid holdup calculation into two parts.36 CHAPTER 3. The horizontal . T ransitionF low occurs when CL ≥ 0.3. EL (θ). 2 3 3.12: Intermittent Flow Regime (retrieved from[4]). EL (0). is determined.2 Hydrostatic Pressure Diﬀerence Once the ﬂow pattern has been determined. this horizontal holdup is corrected for inclined ﬂow to obtain the actual holdup. DistributedF low occurs when CL < 0.1. MODEL FORMULATION Figure 3. the liquid holdup for horizontal ﬂow. First.13: Distributed Flow Regime (retrieved from[4]). the liquid holdup is then calculated. Afterwards.01 and L∗ < F rm < L∗ . 1 4 Figure 3.
4846 0.8θ) 3 β for all type of ﬂow pattern is β = (1 − CL )ln 0.065CL EL (0) ← F r0.0609 m end if if strcmp(F lowpattern. Therefore.98CL EL (0) ← F r0.98CL EL (0)Segregated ← F r0.8θ) − sin3 (1.845CL 0.5824 1.29) (3.3. TWO PHASE FLOW CORRELATIONS 37 holdup must be EL (0) ≥ CL .3692 0.0173 F rm m EL (0)intermittent ← 15: EL (0)T ransition ← AEL (0)Segregated + BEL (0)Intermittent 16: end if 17: end if Once the horizontal insitu liquid volume fraction is determined.30) (3.845CL EL (0) ← F r0. the horizontal holdup is set to EL (0) = CL .4846 0.0978 CL F rm (3.70Nvl 0. The expression used to calculate the horizontal holdup changes per ﬂow pattern group as follows in MatLab: 1: 2: 3: 4: 5: 6: 7: 8: 9: 10: 11: 12: 13: 14: if strcmp(F lowpattern.0868 0.5351 0.1244 4.3. Distributed ) then 0. in the event that EL (0) < CL .0868 m else if strcmp(F lowpattern.5351 0.0173 m end if if strcmp(F lowpattern. Intermittent ) then 0. the actual liquid volume fraction is obtained by correcting EL (0) by an inclination factor B(θ): EL (θ) = B(θ)EL (0) 1 B(θ) = 1 + β sin(1. Segregated ) then 0. T ransition ) then L∗ −F rm 3 A ← L∗ −L∗ 3 2 B ←1−A 0.31) .
the mixture density ρm is obtained.8725(ln(Y ))2 + 0. MODEL FORMULATION where Nv l = 1. we utilize the Moody Friction factor calculated using the Colebrook equation. ρL gσ 1 4 and β must always be ≥ 0.1.938Vsl of β is obtained.33) Based on experimental data.3 Frictional Pressure Loss In order to calculate frictional losses. in turn.0523 + 3. is used to calculate the pressure change due to the hydrostatic head of the vertical component of the pipe ∆PHH = ρm gLsin(θ) 144gc (3.3.34) (3.32) 3. Beggs and Brill presented a correlation for the ratio of the twophase friction factor ftp and the normalizing (noslip) friction factor resulting in the following exponential equation: ftp = fN S eS The value of S depends on the noslip and the actual liquid holdup: S= ln(Y ) −0.35) where Y = .38 CHAPTER 3. Therefore if a negative value Once the actual liquid holdup EL (θ) is calculated. β=0.182(ln(Y )) − 0. Mixture density. To determine fN S . For this purpose.01853(ln(Y ))4 CL EL (θ)2 (3. the noslip Reynolds number is used: ReN S = ρN S Vm D µN S (3. a normalizing friction factor fN S is used.
3. Due to the large volume of steam(gas) as compared to the water that is present in most steam injection wells.2.2Y − 1. [29].38) . [29] mechanistically based pressure drop correlation for downwards steam ﬂow.3.4 ρL 62.2) else ln(Y ) S ← −0.01853(ln(Y ))4 end if Finally.1 Flow Pattern Determination In Figure 3. Those modiﬁcations are included in the bubble ﬂow and in the slug ﬂow patterns. TWO PHASE FLOW CORRELATIONS 39 1: 2: 3: 4: 5: if Y ≥ 1 and Y ≤ 1. As Fontanilla stated for Aziz et al.0764 1 3 Nx = VSg 72 σL ρL 62. the annular mist ﬂow pattern is very common.8725(ln(Y ))2 +0.3.2 Aziz. Govier and Fogarasi Model Fontanilla[19] investigated applicability of the Aziz et al.3.37) Ny = VSL 72 σL (3.4 1 4 1 4 (3.182(ln(Y ))−0. 3.0523+3. ρg 0.36) 3.14 we can see the ﬂowpattern map for diﬀerent ﬂow types. this correlation that was strictly developed for upward vertical ﬂow needs some modiﬁcation in order to apply to downward ﬂow.2 then S ← ln(2. the expression for pressure loss due to friction is: ∆Pf = 2 2ftp Vm ρN S L 144gc (3.
42) where vbf is the rise velocity of small gas bubbles in a ﬂowing liquid. ρg is in pounds per cubic feet.152 (3.51(100Ny )0. In downward ﬂow. vS is in feet per second.8Ny N3 = 70(100Ny )−0. In the Figure 3. (retrieved from[14]). Liquid holdup for bubble ﬂow is calculated from HL = 1 − vSg vbf (3. Bubble Flow The bubble ﬂow pattern is characterized by small bubbles of steam(gas) dispersed in a continuous water phase.40) (3.39) (3.172 N2 = 8. the diﬀerence in densities of the two phases causes the bubbles to travel at a velocity lesser than the average velocity of the mixture.40 CHAPTER 3.41) where vSL is in ft per second. Bubble ﬂow exists if Nx < N1 . ρL is in pounds per cubic feet and σL is in dynes per centimeter.14: Flow Pattern map for Aziz et al.6 + 3.14 we have several lines that represents ﬂowpattern transitions and they are deﬁned as: N1 = 0. MODEL FORMULATION Figure 3. This velocity is .
The vbs term is predicted from σL g(ρL − ρg ) vbs = 1.2vm − vbs (3.43) where the ﬁrst term is the approximate velocity of the ﬂuid mixture.44) The frictional component of the pressure gradient is determined as 2 f ρs vm ∂p = ∂Z 2d (3. . ρs = ρL HL + ρg (1 − HL ) NRe = ρ L vm d µL (3. TWO PHASE FLOW CORRELATIONS 41 predicted for downward ﬂow as follow vbf = 1.47) The acceleration component of the pressure gradient is considered to be negligible for bubble ﬂow.3.34) and friction factor from Moody Friction factor calculation using the Colebrook equation[27].41 ρ2 L 1 4 (3.3. and vbs is the rise velocity of a continuous swarm of small bubbles in a static liquid column.45) where ρs is determined from equation (4. accounting for the nonuniform velocity and bubble concentration proﬁles across the cross section.46) (3.
[29] state that gd(ρL − ρg ) ρL vbs = C where C was given by Wallis as (3.5 for Ny ≥ 4.37−NE m ) (3.(4.345 1 − e(−0.51) if Nv ≥ 250 then m ← 10 else if 250 > Nv > 18 then −0.30) and Eq. Aziz et al.49) gd2 (ρL − ρg ) σL d3 gρL (ρL − ρg ) µL (3.37).50) (3. however. For slug ﬂow.(4.42 CHAPTER 3.029Nv ) and NE = Nv = and m is determined as 1: 2: 3: 4: 5: 6: 7: 8: 9: 10: 1 − e( 3. the bubblerise velocity in a static liquid column is based on large bubble. MODEL FORMULATION Slug Flow Slug ﬂow exists if N1 < Nx < N2 for Ny < 4 or N1 < Nx < 26.35 m ← 69Nv end if if Nv ≤ 25 then m ← 25 end if end if .48) C = 0. The liquid holdup for slug ﬂow is also calculated from Eq.
method if the ﬂow pattern map in Figure 3. linear interpolation is performed. In Figure 4. Aziz et al. TWO PHASE FLOW CORRELATIONS 43 The friction pressuregradient component for slug ﬂow is determined from dp dZ = f 2 f ρL HL vm 2d (3. the transition region does not exist for Ny > 4.3. The Reynolds number is given as NRe = f ρL vm d µL (3. Transition Flow The transition region exists when N2 < Nx < N3 for Ny < 4.14 was replaced with the Duns and Ros map.53) The acceleration pressuregradient component was considered negligible for slug ﬂow. dp =A dZ where A = 3.5 for Ny > 4.54) mist Modiﬁcations AlNajjar and AlSoof [5] showed that improved results could be obtained with Aziz et al.3.3.2. When the transition region is predicted. the pressure gradients must be calculated with both the slugﬂow and mistﬂow equations.2 N3 −Nx N3 −N2 dp dZ + (1 − A) slug dp dZ (3.8. Their conclusion was based on a comparison of the predicted and measured .52) The friction factor is obtained from a Moody friction factor and the Colebrook equation[27]. Mist Flow Mist ﬂow exists when Nx > N3 for Ny < 4 or Nx > 26. To obtain the pressure gradient.[29] recommended the Duns and Ros [28] mistﬂow method be used to calculate pressure gradient for this ﬂow pattern.
MODEL FORMULATION pressure drops for 80 tests on 15 ﬂowing wells in Iraq.44 CHAPTER 3. .
however.1) where Ptotal is the total pressure of the gas phase.Chapter 4 Eﬀect of NonCondensable Gas (N2) In our calculations we used N2 as the noncondensable gas. N2 is assumed not to carry heat. 45 . Neglecting the enthalpy of N2 is a good approximation because steam enthalpy is several orders of magnitude larger in comparison to N2 . temperature. pressure and as well as heat loss during the steam injection. We explore the partial pressure eﬀect of N2 on downhole steam quality. provides partial pressure during steam injection. Addition of noncondensable gas into our system makes our calculation a little complex [10]. N2 . so contribution to the mixture enthalpy is zero. The pressure of the vapor phase is now a sum of steam partial pressure and gas partial pressure: Ptotal = Psteam + PN2 (5. We assume N2 only exists in the vapor phase with steam and has the same temperature as the steam. Psteam is the partial pressure of the steam and PN2 is the partial pressure of the N2 .
46 CHAPTER 4. steam quality.5) Suppose we have a system with given injection rate.3) where x is the liquid mole fraction of the components. we ﬁnd the total pressure as Ptotal = Psteam Psteam = ysteam 1 − yN2 (5. So. 5. time and mol fraction of N2 .3 and y2 = yN2 + ysteam = 1. EFFECT OF NONCONDENSABLE GAS (N2 ) The total mol fraction of the liquid and gas phase is xi = 1.7) Having the total mass of the steam in our system.0 yields that ysteam = 1 − yN2 . steam temperature or pressure (one determines the other). xi = 1. we ﬁnd the total moles of the steam . y is the vapor mole fraction of the components. With a given undilited steam pressure or temperature (one determines other).0 yi = 1.6) Once we ﬁnd total pressure we continue to ﬁnd other input parameters in our calculation. Total mass of the steam is found as WmX = W m ∗ t ∗ x (5.4) (5. Because we have only liquid water in our system inequlibrium with steam and N2 .0 is always equal to xwater . mole fraction of steam is found from Eq.2) (5. Calculation of the partial pressure of the gas components is PN2 = yN2 ∗ Ptotal Psteam = ysteam ∗ Ptotal (5.0 (5.
The reason we calculated mole of the N2 is because it is not changed entire process. Another important assumption is that the viscosity of mixture of N2 and steam is the as same as the viscosity of the undiluted steam.02 (5.9.9) Because ysteam is known and yN2 is known. M oleN2 is easily found using Eq. there is only one unknown in Eq. When this is the case the gas and steam act as perfect gases.9. cases occur where the gas partial pressure is signiﬁcant compared to the steam partial pressure. R is the universal gas constant 8.3. Total volume of the steam and N2 is obtained using ideal gas law as Vsteam = M oleX ∗ R ∗ Tsteam Psteam@initial (5. The equation for viscosity is given by Eq.15.10) where P is the total pressure of the system in kilo pascal.23.47 as M oleX = WmX ∗ 454 18. 5. n is the mole fraction of the component.8) . Meaning that mass of the N2 is constant al the time. K =o C + 273. and it is going to increase due to the condensation of steam.8) With known steam mole fraction. we ﬁnd how many moles of N2 is in the system as ysteam = M oleX M oleX + M oleN2 (5.7) VN2 = M oleN2 ∗ R ∗ Tsteam Psteam@initial (5. The only thing is going to change in each interval is that mol fraction of the N2 . 5. Thus we use ideal gas law PV = n ∗ R ∗ T (5. We assume that in two phase ﬂow.314 J/Kmol and T is the temperature in Kelvin.
The density of the gas mixture and liquid mixture are used in our calculation. That is. the overall volume is found by applying Amagat’s law. potentially more accurate.48 CHAPTER 4. . In each segment of the wellbore. the volumes of each component add ideally and there is no volume change upon mixing. In either case. method is to obtain the steam molar volume from the steam tables. it is easy to ﬁnd density of the total gas mixture. this calculation has to be repeated until steam reaches the reservoir. EFFECT OF NONCONDENSABLE GAS (N2 ) Once steam volume and N2 volume are obtained. Another.
The fourth snapshot is the output tables both onshore and oﬀshore environments. we provide solutions for the addition of noncondensable gas N2 for both onshore and oﬀshore environments. The user can choose diﬀerent twophase correlations and get the results as well as choose previous calculation results for postprocessing. The theory behind calcualtions is well described in preceding chapters and the Appendix.Chapter 5 Graphical User Interface (GUI) Our GUI provides several inputs parameters and options that are useful in the calculations. The second snapshot is about calculations for onshore environment with insulated and without insulated tubing. The ﬁfth snapshot is about the postprocessing for viewing the results in ﬁgures for steam quality. Also. steam temperature and heat loss along the wellbore. All 49 . This chapter shows how we get the results using our GUI. The overall heat transfer coeﬃcient is implemented both following Willhite’s approach[49] and Ramey’s time function[42]. steam pressure. This is similar to Fontanilla [19]. GUI for the Main program The GUI is demonstrated with snapshot of the program. The ﬁrst snapshot gives the background for the program. The third snapshot is the input parameters for the oﬀshore environment.
it is seen that there are several input parameters for speciﬁcally steam injection parameters for onshore cases. is postprocessing the data that obtained from calculations either onshore or oﬀshore is designed speciﬁcally to visualize the results and to save the ﬁgures or delete it.1. User can choose diﬀerent insulation materials and twophase ﬂow correlations. . In third ﬁgure that is Figure 5. Figure 5.4. Last one is the Figure 5. results are shown this part and user able to save those datas to excel ﬁles or clear them. First snapshot of the program is shown Figure 5. after the calculation is done either for onshore or oﬀshore cases. User also able to choose insulation materials and two phase ﬂow correlation to calculate steam temperature. Figure 5.2 is the similar input parameters except riser radius and sea temperature inputs.1: User interface developed GUI for onshore calculations.3. steam pressure. GRAPHICAL USER INTERFACE (GUI) calculations take into account twophase ﬂow correlations.50 CHAPTER 5. steam quality and heat loss values.
.2: User interface developed GUI for oﬀshore calculations. Figure 5.51 Figure 5.3: User interface developed GUI for both onshore and oﬀshore results.
GRAPHICAL USER INTERFACE (GUI) Figure 5.4: User interface developed GUI postprocessing for both onshore and oﬀshore results. .52 CHAPTER 5.
It is shown that our results match with Prats’s [41] results for surface lines. And then results are extended to 2 ﬁeld cases [11]. 6. diﬀerent depth and diﬀerent percentage of N2 addition in order to see the sensitivity of steam temperature. noncondensable gas addition calculations are done with diﬀerent injection rate. 53 . so we take input values from Prats [41] and modify the calculations to illustrate the hole of insulation. steam pressure. steam quality and heat loss.1 Examples for heat loss calculation Examples with known solutions are presented for surface line and onshore calcualtions with insulated or without insulated tubing to verify the new code. We compared our results with Prats’s [41] results. our results are a little higher than his results. steam quality. There is no example for oﬀshore heat loss calculation. sea and sea ﬂoor to reservoir heat loss calculations are presented using limited data provided by Prats [41]. This part of results are only considered for heat loss aspects. Moreover. steam temperature. surface line. With this we found his minor mistake for onshore injection well and reported here. For injection wells.Chapter 6 Results and Comparisons In this part of the study.
3. on the other side when you use aerogel you would consume 16 bbl for 100 ft length surface pipe. When you use calcium silicate you would consume 100 bbl.095*109 BTU when the pipe is bare.0*106 BTU. the reduction in yearly heat losses resulting from insulating the pipe amounts to more than 1000 bbl of fuel for a 100ft length of pipe.1. the insulation reduces heat losses by a factor of about 50 when using calcium silicate.2 and plotted into Figure 6. the amount of heat lost from 100ft length of pipe over a period of 1 year is Ql =1.” The average yearly temperature is 60 o F . Prats [41] asks ”ﬁnd the steady state heat loss per year per 100 ft of pipe when the pipe is (1) is insulated with 3 in of calcium silicate and (2) not insulated .e. and the prevailing winds have an average velocity of 20 mph normal to the injection line. The input parameters are listed in Table 6.2 is also applied here in order ﬁnd the heat loss for without insulation case.6 18vw rins rins hf c = (6. for black aerogel heat loss is Ql =2. Eq.6 0.1581*108 BTU for the calcium silicate (highest heat loss).i.1 for a several pipe sizes and temperatures. radiation heat losses are important.1) With insulation. N80 pipe at a rate of 229 B/D. 0. Therefore. RESULTS AND COMPARISONS 6. the over all speciﬁc thermal resistance is calculated from Eq. however if you use aerogel it would be 280.1 from Prats Steam at 550 o F is injected through 4in. Accordingly. Because one barrel of oil is roughly 6. Thus. Without insulation and at a surface temperature near 550 o F . bare.1 Example 10. The sum of the coeﬃcients of heat transfer due to the radiation and free(or natural) convection for a horizontal pipe is given Table 3.0302*107 BTU (lowest heat loss).1 and for without insulation plotted into Figure 6. When we think about deeper wells using insulation material . the amount of heat lost from a 100ft length of pipe over a period of 1 year is: Ql =6.. 3.54 CHAPTER 6.1.2.
”SeawithIns” is the oﬀshore case with insulated tubing.1. Ql =1. Without insulation. Prats [41] stated that the coeﬃcient of heat transfer due to radiation and forced convection is estimated to be 330 BT U/f t2 −D −o F .1 with insulations.1. In Example 10.21*109 BTU.6. Prats [41] used calcium silicate as the insulation material and get the heat loss as.095*109 BTU as shown in Figure 6. over the year. This shows that the real heat loss from bare tubing is Ql =6. Table 6. In Table 6. Abbreviations are explained as ”SLwithIns” is the surface line with insulated tubing. and ”2ResWithoutIns” is the sea ﬂoor to reservoir case without insulated tubing. and it causes little deviation from the exact results. he used 110 BT U/f t2 −D −o F and get the heat loss results over the one year period is Ql =6.1581*108 BTU and it is consistent with Prats’s example as shown in Figure 6.16*108 BTU.1 is used when we calculate the heat losses for the uninsulated case. ”SeawitouthIns” is the oﬀshore case without insulated tubing. . EXAMPLES FOR HEAT LOSS CALCULATION 55 should be considered especially aerogel. ”2ResWithIns” is the sea ﬂoor to reservoir case with insulated tubing. ”SLwithoutIns” is the surface line without insulated tubing. In our calculation we ﬁnd this minor mistake from Prats calculation. however. Our result for using calcium silicate is Ql =1.2. Still those results match well.1. there are several column headings.
9 0.6 0.1458 0.9 600 100 0.166/24 365*24 100 0 0 0 0 0 0 550 60 0.1478 0.2292 0.9 0.166/24 21*24 164 0.1458 0.9 CHAPTER 6.1667 0.2292 0 0 0 0 0 0 24 0. Parameters SLwithIns SLwithoutIns SeaWithIns SeaWithoutIns 2ResWithIns 2ResWithoutIns Tsteam Taverage rti rto rins rriserin rriserout hf hpi hpo λpipe λcem λins time P ipeLength riserin riserout rto riserin riserout rto 550 60 0.1458 No Insulation 0.1458 No Insulation 0 0 0 0 0 0 24 No Insulation 21*24 1000 0 0 0 0.75 0 0 0 25 24 0.166/24 21*24 1000 0 0 0 0.14 0.1478 0.9 0 0 0 600 100 0.9 0. RESULTS AND COMPARISONS .1: Input parameters from Prats [41] as used for diﬀerent example calculations.1667 No Insulation 0 0 2000 ∞ 2000 25 0 No Insulation 365*24 100 0 0 0 0 0 0 600 70 0.9 0.9 0.14 0.9 0.9 0.41467 0 0 2000 ∞ 2000 25 0 0.6 0.14 0.75 0 0 0 25 24 No Insulation 21*24 164 0.9 0 0 0 600 70 0.14 0.56 Table 6.9 0.
2 69.8 580 680 780 880 980 127 124 120 116 112 111 108 149 146 141 137 134 132 129 174 171 166 162 158 156 153 202 198 194 189 186 184 180 234 230 225 221 217 215 212 1080 269 265 260 256 252 250 247 1180 307 304 299 294 290 289 286 1280 352 348 343 338 334 332 329 57 .50 1.3 62.2 42.00 12.00 8. Diameter 130 0.8 49.4 180 59.7 65.0 100.4 72.9 48.4 48.0 39.9 76.4 71.1 72.2: Radiationnatural convection coeﬃcient of heat transfer.5 93.8 330 81.00 2.0 96.1 54.8 75.8 51.6.0 86.2 46.5 68.9 83.9 88.5 51.2 63.1 91.00 50.1.3 79.4 61.7 46.6 280 74.00 4.6 60.5 57.4 67.7 480 107.3 44.8 70.4 65. EXAMPLES FOR HEAT LOSS CALCULATION Table 6.2 41.8 79.0 58.3 230 66.1 53.0 104.00 24.3 380 90.0 75.5 57.
2: Surface Heat Loss calculation without insulation.58 CHAPTER 6.1: Surface lines heat loss calculation with six diﬀerent insulation materials. RESULTS AND COMPARISONS Figure 6. Figure 6. .
1.3: Heat loss from sea level to sea ﬂoor with six diﬀerent insulations. EXAMPLES FOR HEAT LOSS CALCULATION 59 6. Figure 6.1.6.4: Heat loss from sea level to sea ﬂoor without insulation. .2 Example for Oﬀshore Figure 6.
the insulation being held in place and sealed from accidental entry of liquids in the annulus by a very thin sheath of aluminum. The temperature at the surface of the insulation (Tins ) and that at the inner radius of the casing (Tci ).5lbm/ft N80 casing. Estimate the rate of heat loss 21 days after steam injection started.5in.87 ∗ 104 ) BTU gives four times more heat loss than black aerogel (0.3 Example 10. 6. Radiation is sensitive to the temperature levels and emissivities (∈) of the surfaces. tubing set on a packer in 9 5/8 in. aﬀect the radiation heat losses across the annular space between the insulated casing.60 CHAPTER 6. as well as the casing temperature. For instance using calcium silicate (2. of calcium silicate.1.1.. Heat transfer across the gas ﬁlled annulus is by radiation and natural convection.1 ∗ 105 ) BTU case also gives about 8 times more heat loss than using lowest thermal conductivity insulation materials. There is no altered zone near the boreholes. In addition. It is clear that for 21 days period heat loss are substantial for uninsulated wells. The annulus contains a stagnant gas at zero gage pressure at wellhead. the bare tubing (2. together with the emissivities at these surfaces (∈ins ) and (∈ci ). hole. A temperature survey in the well indicates a mean surface temperature of 100 o F over the 1000ft depth. The calculation procedure is well explained in the Prats book[41]. steam injection for oﬀshore heat loss is calculated with insulated material and uninsulated cases. Prats [41] assumed that for shallow reservoir temperature change does not vary . Sea temperature is taken as 70 o F for 164 f t water depth. Heat loss estimation for 21 days of injection shows that the greatest the thermal conductivity values the greatest the heat losses observed. The tubing is insulated with 1 in. RESULTS AND COMPARISONS Under 600 o F . This is relatively shallow. and the casing is cemented to surface in a 12in.2 from Prats Steam at 600 o F is injected down 3. 53.75 ∗ 104 ) BTU. Input parameters are taken from Table 6.
In our calculation we get Ql =4.5. It is because of assuming temperature of the ﬂuid does not change with depth for shallow reservoirs. He also stated that when the steam injection happens.1. This is good approach in terms of calculation and simplicity. Discussion of heat loss so far. This is why using a single value of Rh may provide a close enough estimate of heat losses and wellbore temperatures. the daily heat loss from such an insulated well corresponds to less than 1 bbl of fuel.2 from Prats [41] got heat loss Ql =4. It is because of several input parameters that diﬀers somewhat. shallow reservoir temperature stays constant. considers the entire well as a unit. EXAMPLES FOR HEAT LOSS CALCULATION 61 much with depth such as 1000 ft.6*106 BTU that is slightly diﬀerent than Prats results that is shown in Figure 6. In our program validation we do not assume this and we take into account temperature change along the wellbore.6. Example 10. .0*106 BTU/bbl of fuel.46*106 BTU for 21 days that corresponds to an equivalent energy content of 6.
2 from Prats [41]. .62 CHAPTER 6.5: Heat loss calculation using diﬀerent insulation materials based on Example 10. Figure 6.2 from Prats[41].6: Heat loss calculation without using insulation materials based on Example 10. RESULTS AND COMPARISONS Figure 6.
The Martha Bigpond well data is retrieved from the paper.104166667 No Insulation 0.1875 0.6.9 0.9 0.9 0.7 117 Wm x pwh Tm Depth Pan t There are 2 ﬁeld cases where steam temperature is reported in the literature [11].2 0.104166667 No Insulation 0. PROGRAM VALIDATION 63 6. Fontanilla and Aziz [20] .8 250 50 1600 14. Input Parameters Field Data 1 Field Data 2 Dti rto rins rci rco rh ke αE kcem to ci EART H 0.94 4640 0.0286 0.177 0.2 0.6 1 0.9 0.94 4850 0.2 Program Validation Table 6.1875 0.166666667 0.2.3: Field data parameters for ﬁeld data 1 and ﬁeld data 2 [19].8 250 50 1600 14.177 0.0286 0.6 1 0.7 71 0.166666667 0.
. we can see the heat loss during steam injection is almost same values for the four diﬀerent correlations of multiphase ﬂow. RESULTS AND COMPARISONS referred to the same ﬁeld data from Bleakley’s paper [11] as ﬁeld data 1 and 2.7 and 6. the Beggs and Brill over predicts the quality values that Fontanilla got for ﬁeld data 1 in Figure 6. [29].10. Input parameters are tabulated in Table 6. It may be because of applying diﬀerent correlations to get saturated steam properties during calculations. we got promising results using the Beggs and Brill [13. When we look at the steam quality versus depth. The other three correlations also get good agreement with Fontanilla’s results as shown in Figure 6. 14] and Aziz et al.9.8.64 CHAPTER 6.3. Hasan and Kabir model [23. 14] approach for multiphase ﬂow with ﬁeld data 1. 40] and the Drift Flux model [21] are applied in our calculations. In Figure 6. These cases are to referred in our calculations as ﬁeld data 1 and ﬁeld data 2 as well. After applying the modiﬁed Fontanilla[20] approach with modiﬁed correlations. The other three correlations that we implemented have almost the same results. Besides Beggs and Brill [13.
PROGRAM VALIDATION 65 Steam Temperature in 355 0 360 365 370 375 380 385 o F 390 395 400 405 200 400 600 Field Data 1 Fontanilla Aziz.8: Comparison of steam pressure with ﬁeld data 1 and twophase correlations.2. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 800 1000 1200 1400 150 1600 160 166 168 Figure 6.7: Comparison of steam temperature with ﬁeld data 1 and twophase correlations. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux Depth(ft) 800 1000 1200 1400 1600 358 o F 364 o F 368 o F371 o F Figure 6. . 0 140 160 Steam Pressure in psi 180 200 220 240 200 400 600 Depth(ft) Field Data 1 Fontanilla Aziz.6.
Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 600 Depth(ft) 800 1000 1200 1400 1600 0. RESULTS AND COMPARISONS 0.72 0.66 CHAPTER 6.7 0 0.77 0.79 0.10: Calculated heat loss calculation with insulated tubing based on ﬁeld data 1.8 200 400 Fontanilla Aziz.9: Calculated steam quality with diﬀerent twophase correlations based on ﬁeld data 1.71 0.75 0.78 0. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 600 Depth(ft) 800 1000 1200 1400 1600 x 10 5 Figure 6.703 0. .74 0.745 Figure 6.76 0.73 Steam Quality 0. 0 0 1 2 3 Heat Loss in BTU 4 5 6 7 8 9 10 200 400 Aziz.
The other three correlations give results similar to Fontanilla’s result. our calculation shows a little overprediction of values for steam quality as shown on Figure 6. . Again. Although. 355 0 360 365 Steam Temperature in 370 375 380 385 o F 390 395 400 405 200 400 600 Depth(ft) Field Data 2 Fontanilla Aziz. we got good results for steam temperature and pressure values for Beggs and Brill.11 and 6. PROGRAM VALIDATION 67 Figure 6.12 also show results after we run our simulator for ﬁeld test data 2.2. we can see that the modiﬁed Beggs and Brill method gives a better result than the Fontanillla approach compared to the ﬁeld data 1.6.11: Comparison of steam temperature with ﬁeld data 2 and twophase correlations.13 from Fontanilla’s result. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 800 1000 1200 1400 1600 Figure 6.
13: Calculated steam quality with diﬀerent twophase correlations based on ﬁeld data 2. .72 Steam Quality 0.78 0.68 CHAPTER 6.76 0.7 0. 0.12: Comparison of steam pressure with ﬁeld data 2 and twophase correlations. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6.68 0 0.74 0.8 200 400 Fontanilla Aziz. RESULTS AND COMPARISONS 140 0 150 160 170 Steam Pressure in psi 180 190 200 210 220 230 240 250 200 400 600 Depth(ft) Field Data 2 Fontanilla Aziz. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 800 1000 1200 1400 1600 Figure 6.
.15: Comparison of our model with Fontanilla’s model steam temperature with ﬁeld data 1. 340 0 350 360 Steam Temperature in 370 380 o F 390 400 410 200 Field Data 1 Our Model(with Aziz) Our Model(with Beggs&Brill) Fontanilla Model(with Aziz) Fontanilla Model(with Beggs&Brill) 400 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6.2. Govier and Fogarasi Beggs & Brill Hasan & Kabir DriftFlux 600 Depth(ft) 800 1000 1200 1400 1600 x 10 5 Figure 6. PROGRAM VALIDATION 69 0 0 2 4 Heat Loss in BTU 6 8 10 12 14 16 200 400 Aziz.14: Calculated heat loss calculation with insulated tubing based on ﬁeld data 2.6.
77 0.75 0.8 200 400 Our Model(with Aziz) Our Model(with Beggs&Brill) Fontanilla Model(with Aziz) Fontanilla Model(with Beggs&Brill) 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6.79 0.16: Comparison of our model with Fontanilla’s model steam pressure with ﬁeld data 1.71 0.72 0.74 0.7 0 0.78 0.17: Comparison of our model with Fontanilla’s model steam quality with ﬁeld data 1. RESULTS AND COMPARISONS 120 0 140 160 Steam Pressure in psi 180 200 220 240 260 200 400 Field Data 1 Our Model(with Aziz) Our Model(with Beggs&Brill) Fontanilla Model(with Aziz) Fontanilla Model(with Beggs&Brill) 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6.70 CHAPTER 6. Steam quality 0.76 0.73 0. When we look at the results that we got for both ﬁeld data 1 and 2 cases. The ﬁgures above . we had an opportunity to compare the results with Fontanilla’s approach.
17. PROGRAM VALIDATION 71 for the ﬁeld data 1.15.18: Comparison of our model with Fontanilla’s model steam temperature with ﬁeld data 2. although we are able to calculate steam quality with given injection rate and input parameters. Beggs and Brill results diﬀer quite a bit between the two methods. In Figure 6. Looking at the results. Therefore we only compare the results we have in our model and Fontanilla’s model. On the other hand. model starts converging from the beginning and obtains similar values. . 330 0 340 350 Steam Temperature in 360 370 o F 380 390 400 410 200 Field Data 2 Our Model(with Aziz) Our Model(with Beggs&Brill) Fontanilla Model(with Aziz) Fontanilla Model(with Beggs&Brill) 400 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6. On the other hand Fontanilla’s Beggs and Brill implementation has less accurate results than ours. in Figure 6. has good agreement with Fontanilla’s Beggs and Brill results.15. It comes to check steam quality. and our implementation of Beggs and Brill correlations starts converging to the real data when we reach the bottom of the wellbore in Figure 6. Fontanilla’s Aziz et al.16. calculations give almost perfect results with the ﬁeld data until 1000 ft.6. in the literature we could not ﬁnd the steam quality data. we can also see the same trend with Figure 6.2. We can conclude the comparison saying that the Aziz et al. Our model using Aziz et al. Reuslts then deviate from the real values and go out of the range.
68 0 0.20: Comparison of our model with Fontanilla’s model steam quality with ﬁeld data 2. Steam quality 0. In Figure 6.18 .6.74 0.78 0. In the ﬁeld data 2.7 0. we got the results and plot them as shown in Figures 6.8 200 400 Our Model(with Aziz) Our Model(with Beggs&Brill) Fontanilla Model(with Aziz) Fontanilla Model(with Beggs&Brill) 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6.19: Comparison of our model with Fontanilla’s model steam pressure with ﬁeld data 2.20.18 our model with Beggs and Brill correlations gives very good agreement . RESULTS AND COMPARISONS 100 0 120 140 Steam Pressure in psi 160 180 200 220 240 260 200 400 Field Data 2 Our Model(with Aziz) Our Model(with Beggs&Brill) Fontanilla Model(with Aziz) Fontanilla Model(with Beggs&Brill) 600 Depth(ft) 800 1000 1200 1400 1600 Figure 6.72 0.76 0.72 CHAPTER 6.
For the oﬀshore part. the Beggs and Brill twophase ﬂow correlation model’s temperature proﬁle or pressure proﬁle drops dramatically. In contrast. three other twophase correlations converge to the same values both on temperature proﬁle . The Aziz et al. and various twophase correlations are used. time is 1 year and injection rate 4850 lbm/hr similar to the onshore. On the other hand. approach initially gives great results and captures almost exact values for several hundreds ft then deviates substantially from the reality. The last comparison is steam quality. results better match ﬁeld data as compared to Fontanilla’s Aziz et al. approach. the Aziz et al. ONSHORE ENVIRONMENTS 73 with the real data and converges to the exact value.20. We can also conclude with saying that we improved Fontanilla’s approach using several twophase correaltions.6. however. This trends continue in Figure 6.3. oﬀshore and noncondensable gas cases.19 with pressure values. We can say that our model with Beggs and Brill twophase ﬂow correlations has less quality drop than Fontanilla’s Begss and Brill model in Figure 6. approach has similar steam quality values in both models. Three diﬀerent steam injection temperatures and two diﬀerent depths are considered for the onshore environment. Fontanilla’s method with the Beggs and Brill approach gives less accurate results than our model. Steam temperature. steam pressure. 6. method.3 Onshore environments For both onshore and oﬀshore environments with insulated or uninsulated tubing input parameters are used for ﬁeld data 1. steam quality and heat loss values are investigated. Results in Figures 6. The surrounding temperature. When the steam temperature is 400 o F .22 are insulated by black aerogel.21 and 6. Fontanilla’s Aziz et al. is taken as 122 o F . however. ﬁeld data 1 values are used as an input and exception is also used here with additional input for oﬀshore sea temperature is taken as 70 o F . In our model.
the behavior of the Beggs and Brill model when the injection temperature increased to 600 o F gives less temperature drop than all the other three correlations. Although the oﬀshore depth of 200 f t is relatively shallow and greater pressure drop and temperaure drop for small interval does not really eﬀect the behavior trend of the steam properties at downhole conditions. therefore this explanation will help to understand those ﬁgures as well. it gives a little greater value than other correlations.27 show results. Figures 6. Without insulation. Figures presented in Appendix C also have similar behavior in our example. When the depth is greater than 8000 f t. the greater steam quality and steam properties are obtained in both cases. The steam quality values with insulated tubing when compared to uninsulated tubing cases diﬀer signiﬁcantly.74 CHAPTER 6. there is greater heat loss and quality decreases to a greater extent as compared to the insulated case. with increasing injection temperature overpredicts the steam quality values. RESULTS AND COMPARISONS and pressure proﬁle. . The greatest heat loss is obtained using calcium silicate which has greatest thermal conductivity value among the six diﬀerent insulations. Similar trends are obtained during the steam injection in an oﬀshore environment as well. Figure 6.26 show that with diﬀerent temperature and depth all the ﬂow regimes have similar values in steam temperature. the Beggs and Brill twophase ﬂow correlation model’s temperature or pressure proﬁle drops less than the other correlations until 7500 f t then converge to the others.23 and Figure 6. The diﬀerences between insulated and uninsulated tubing is the heat loss and steam quality values. Interestingly. and steam quality comparing with insulated case. It is also observed that.24 and 6.25 and 6.28 give the information that increasing injection steam temperature causes greater heat loss when comparing 500 o F and 600 o F cases. steam pressure. The greater the steam temperature. When we increase injection temperature to 500 o F . the Beggs and Brill model. Figures 6.
3. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel.1 Examples with Insulation Materials U singBlackAerogelλBA = 0. 1 year. Figure 6. . ONSHORE ENVIRONMENTS 75 6.6.3.22: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 1 year.0069BT U/(f t − hr −o F ) Figure 6.21: Steam temperature distribution for diﬀerent injection temperature vs depth (ft).
RESULTS AND COMPARISONS Figure 6.24: Heat loss distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. 1 year.23: Steam quality distribution for diﬀerent injection temperature vs depth (ft). 1 year. Figure 6. .76 CHAPTER 6.
ONSHORE ENVIRONMENTS 77 6.25: Steam temperature distribution . 1 year. . Figure 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. 1 year.3.2 Examples without Insulation Materials Figure 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.26: Steam pressure distribution.6.3.
Figure 6. . Tm = 122 o F and injection rate 4850 lbm/hr without insulation. RESULTS AND COMPARISONS Figure 6. 1 year. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.28: Heat loss distribution.78 CHAPTER 6.27: Steam quality distribution.
Figure 6.4 6.0069BT U/(f t − hr −o F ) Figure 6. .4. Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel.6. Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel.29: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). OFFSHORE ENVIRONMENTS 79 6.4. 1 year. 1 year.1 Oﬀshore Environments Examples with Insulation Materials U singBlackAerogelλBA = 0.30: Steam pressure distribution for diﬀerent injection temperature vs depth (ft).
80
CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.31: Steam quality distribution for diﬀerent injection temperature vs depth (ft), 1 year, Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel.
Figure 6.32: Heat loss distribution for diﬀerent injection temperature vs depth (ft), 1 year, Tm = 122 o F and injection rate 4850 lbm/hr for black aerogel.
6.4. OFFSHORE ENVIRONMENTS
81
6.4.2
Examples without Insulation Materials
Figure 6.33: Steam temperature distribution, 1 year, Tm = 122 o F and injection rate 4850 lbm/hr without insulation.
Figure 6.34: Steam pressure distribution, 1 year, Tm = 122 o F and injection rate 4850 lbm/hr without insulation.
82
CHAPTER 6. RESULTS AND COMPARISONS
Figure 6.35: Steam quality distribution, 1 year, Tm = 122 o F and injection rate 4850 lbm/hr without insulation.
Figure 6.36: Heat loss distribution, 1 year, Tm = 122 o F and injection rate 4850 lbm/hr without insulation.
6.4. OFFSHORE ENVIRONMENTS
83
Figure 6.37: Pressure drop distribution and formation pressure (green dots), 1 year, Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel for onshore.
Reservoir pressure is clearly an important factor governing steam injectors. When we know the formation pressure, we can determine whether we can inject steam and at what temperatures in order to heat the reservoir. For this purpose, we could conceive two diﬀerent formation pressures at the same depth. One is smaller than steam pressure one is greater than steam pressure on shown Figure 6.37. If formation pressure is greater than steam pressure, steam can not be injected to the reservoir. Greater the steam pressures are obtained with higher temperatures. Having this advantage, steam can be injected with increasing temperature to the formation.
5 Adding NonCondensable Gas (N2) in an Onshore environment 6.1 Examples with Insulation Materials Figure 6.5. starting from 5 % to 30% mol fraction of the steam. First. the greater N2 . Noncondensable gas is added to the ﬁeld data 1 scenerio an input except 122 o F .38: With changing N2 molar percentage. time is 1 year and injection rate 4850 lbm/hr with modiﬁed Beggs and Brill twophase correlation. RESULTS AND COMPARISONS 6. 1 year. In Figure 6.38. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. diﬀerent N2 molar percentage of steam is added to the steam.84 CHAPTER 6.
so more heat is delivered to the reservoir.5. But. In Figure 6.39. diﬀerent injection rate is studied with 10% by mole N2 added to the steam. it increases steam partial pressure and temperature. however. ADDING NONCONDENSABLE GAS (N2 ) IN AN ONSHORE ENVIRONMENT85 in the system gives more partial pressure to the system. The contribution of partial pressure by steam drops signiﬁcantly and temperature as well. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. 1 year. It has a disadvantage to deliver lower steam temperature to the reservoir in terms of less latent heat addition to the reservoir. Figure 6. In our calculation. Reducing the fraction of N2 in the steam gives greater heat loss and quality drop.39: With changing injection rate. steam quality drops less so the heat loss decreases.6. frictional pressure drops have the greatest contribution .
1 year. Increasing injection rate does not yield good results in downward steam injection operations.40. That may not be good for maintaining the temperature and pressure of the . Frictional pressure drop can be caused by greater injection rate in our case.40: With changing steam quality molar percentage . RESULTS AND COMPARISONS to the pressure drop calculation. Increasing steam quality yields more pressure and temperature drop in existing cases for noncondensable gas. and steam quality decrease downhole.86 CHAPTER 6. The greater the injection rate the greater the pressure drop. Figure 6. diﬀerent steam quality values are conducted in our work with 10% N2 . Figure 6. However. it also yields higher steam quality values when the steam reaches to reservoir with rich quality but poor temperature and pressure values. steam temperature. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel.
41: With changing injection temperature.6. Figure 6. steam pressure. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. smaller pressure drop and steam quality change is obtained with highest steam temperature values. and heat loss. The greatest heat loss is observed with the greatest steam injection temperature at 650 o F . .5. Because of this. heat loss values with diﬀerent steam quality does not vary and have almost same value with diﬀerent injection quality. In Figure 6. Steam temperature values are changed in terms of checking the sensitivity of steam quality. increasing steam temperature is caused by increasing steam pressure. ADDING NONCONDENSABLE GAS (N2 ) IN AN ONSHORE ENVIRONMENT87 steam. In the same ﬁgure.41. steam temperature.
shown in Figure 6.42: With changing injection depth. all the values are corresponded and follow the same distribution except depth.42. For the diﬀerent range of depth. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. after 2500 f t even if steam quality was high. .88 CHAPTER 6. RESULTS AND COMPARISONS Figure 6. injection depth could not be increased. because of the signiﬁcant pressure drop along the tubing. However. The eﬀect of injection depth is analysed by changing the depth.
2 Examples without Insulation Materials Figure 6. 1 year. The two temperature and pressure drops curves have slightly higher values than the insulated case when the N2 mole fraction changes 5 %10%. Steam quality is less than . Another interesting observation is made with comparing to the insulated tubing case. Figure 6.43 has similarities in terms of the greater the mole fraction of N2 we have in the system. However. the greater pressure drop as well as temperature drop. steam quality drops and heat losses values converge to the same value.6. ADDING NONCONDENSABLE GAS (N2 ) IN AN ONSHORE ENVIRONMENT89 6. Same scenerio is also applied for uninsulated tubing as applied to insulated tubing cases.43: With changing N2 molar percentage. The other curves have greater pressure drop compaed to the insulated case.5. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.5.38 and Figure 6.
Using insulated tubing gives advantage instead of bare tubing regarding to less decrease with several parameters.39. but heat loss values almost 10 times greater than without insulated case.44: With changing injection rate. Frictional pressure drop can be caused by more injection rate in our case. Increasing injection rate does not yield good results in downward steam . The greater the injection rate greater the pressure drop. 1 year. RESULTS AND COMPARISONS comparing with insulated tubing.90 CHAPTER 6. In our calculation frictional pressure drops have highest contribution to the pressure drop calculation. steam temperature and and steam quality drop are observed. diﬀerent injection rate is studied with 10% N2 mol fraction of noncondensable gas as shown in Figure 6. In Figure 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. Figure 6.44.
1 year. ADDING NONCONDENSABLE GAS (N2 ) IN AN ONSHORE ENVIRONMENT91 injection operations. Figure 6.45. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. steam pressure. Steam quality and steam pressure or temperature values are inversely proportional. When steam quality increase steam pressure change decrease more as well as steam temperature in existing cases for noncondensable gas.5. Similar trends are seen with insulated tubing except more heat loss and more quality drops are calculated without insulated tubing.6. and vise versa. shows diﬀerent steam quality values are conducted in our work using uninsulated tubing with 10% N2 . steam temperature and greater heat losses. There is a greater decrease of steam quality. it also yields higher steam quality values when the steam .45: With changing steam quality molar percentage . Figure 6. However.
85.6 is 1. RESULTS AND COMPARISONS reaches to reservoir with rich quality but poor temperature and pressure values. . However. For example. the greater the steam quality we have in the system. That may not be good for maintaining the temperature and pressure of the steam. This is the main explanation for the why pressure drop is dramatically for the greatest steam quality value.5 times higher than x = 0. the greater the mixture superﬁcial velocity. the main contribution for frictional pressure drop values is mixture velocity of the components. heat loss values with diﬀerent steam quality does not vary and have almost same values with diﬀerent injection quality. mixture velocity contribution in twophase ﬂow calculations have highest impact on frictional pressure drop.85.6 almost two times. Therefore. I have checked three things : the total density of the mixture. frictional pressure drop for x = 0. The contribution of water to the pressure drop is dependent on the slippage eﬀect value and of course it is going to give little bit higher value than lowest water content on density of the mixture. In this case.92 CHAPTER 6. In the same ﬁgure.85 is almost 2. mixture velocity and frictional pressure drop values for both cases and saw that mixture velocity is higher @ x = 0. so total pressure drop increase more. than @ x = 0.6 case. Mixture density of x = 0. Even if we have greatest liquid water content for the lowest steam quality.2 times higher than x = 0. the pressure drop is the smallest.
1 year.46. steam temperature. Because of this less pressure drop and steam quality change obtained with highest steam temperature values. it is discerned that lowest steam quality. it is seen that the trend of heat loss values is almost 10 times greater as compared to the insulated case. steam pressure and heat loss. In Figure 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. With the insulated case. Steam temperature values are changed in terms of checking the sensitivity of steam quality.5. ADDING NONCONDENSABLE GAS (N2 ) IN AN ONSHORE ENVIRONMENT93 Figure 6. steam pressure. and steam temperature change found with greatest steam injection temperature. Without insulated tubing case.6. The greatest heat loss is observed with the greatest steam injection temperature with 650 o F . Without insulated .46: With changing injection temperature. increasing steam temperature is caused by increasing steam pressure.
94 CHAPTER 6. RESULTS AND COMPARISONS case less steam quality decreases when the temperature 550 o F .47 using insulated tubing. 1 year. a similar trend is shown in Figure 6. . Figure 6. the other higher two values have corresponding values. However. heat loss values are always much greater as compared to the insulated case. even if values seems closer for the two diﬀerent cases. Again. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.47: With changing injection depth. Changing depth under noncondensable gas with using uninsulated tubing is studied by changing the depth ranging from 1600 f t to 2500 f t.
57 have similar trends between onshore cases and oﬀshore cases with uninsulated tubing.46 and Figure 6. on Figure 6.6. quality and heat loss are obtained with onshore environment.55.39 and Figure 6. For example. Figure 6.52 have similar trends between onshore cases and oﬀshore cases with insulated tubing.53.54.42 and Figure 6.38 and Figure 6. ADDING NONCONDENSABLE GAS (N2 ) IN AN OFFSHORE ENVIRONMENT95 6.40 and Figure 6. Figure 6. 6.50. except for initial pressure drop is greater for oﬀshore environment in both insulated and uninsulated cases.45 and Figure 6. This observation continues for uninsulated tubing cases as well. Figure 6. Figure 6.44 and Figure 6.41 and Figure 6.48 heat loss values corresponds on one line.6 Adding NonCondensable Gas (N2) in an Oﬀshore environment Similar observation of steam temperature. and Figure 6. and Figure 6. Figure 6. pressure. For instance. steam quality are similar except heat losses.6. Figure 6.47 and Figure 6.51. steam temperature.1 Examples with Insulation Materials .56.49.43 and Figure 6.48 steam pressure. Figure 6. Figure 6.6.
Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. .48: With changing N2 molar percentage. Figure 6.49: With changing injection rate. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel.96 CHAPTER 6. 1 year. 1 year. RESULTS AND COMPARISONS Figure 6.
Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel. 1 year. ADDING NONCONDENSABLE GAS (N2 ) IN AN OFFSHORE ENVIRONMENT97 Figure 6. Figure 6. .50: With changing steam quality molar percentage .6.6. 1 year.51: With changing injection temperature. Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel.
RESULTS AND COMPARISONS Figure 6. . Tm = 122 o F and injection rate 4850 lbm/hr with using black aerogel.98 CHAPTER 6. 1 year.52: With changing injection depth.
6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. . 1 year.53: With changing N2 molar percentage. 1 year.6.2 Examples without Insulation Materials Figure 6. Figure 6. ADDING NONCONDENSABLE GAS (N2 ) IN AN OFFSHORE ENVIRONMENT99 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.54: With changing injection rate.6.
Tm = 122 o F and injection rate 4850 lbm/hr without insulation. 1 year. RESULTS AND COMPARISONS Figure 6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation. .56: With changing injection temperature. Figure 6.55: With changing steam quality molar percentage . 1 year.100 CHAPTER 6.
57: With changing injection depth. ADDING NONCONDENSABLE GAS (N2 ) IN AN OFFSHORE ENVIRONMENT101 Figure 6. 1 year. .6. Tm = 122 o F and injection rate 4850 lbm/hr without insulation.6.
Solutions take into account pressure drop. solution for heat losses from steam injectors in both onshore and oﬀshore environments. 5) thermolastic insulation and 6) calcium silicate [34] with four diﬀerent correlations for frictional pressure drop in the tubing: Aziz et al. 7. 14.Chapter 7 Summary. 39].1 Summary In this thesis. We used properties that are characteristics of six diﬀerent insulation materials that are 1) black aerogel. Conclusions and Future Work 7. 2) white aerogel. [29]. pressure drop 102 . We made our oﬀshore scenerio calculation taking into account seawater thermo physical properties in order to create a more realistic solution. We compared our solution with two ﬁeld cases and got good agreement. heat loss. Hasan and Kabir [23. 3) ﬁberglass. and multiphase ﬂow. and Drift Flux[21] and gave the results in results section. 4) carbon ﬁber. we have presented an approximate. 37. We improved the solution provided by Fontanilla[19] by using more accurate correlations. but accurate. Beggs and Brill [3. When the thermal conductivity of the insulation material increases. 40] .
although we saw that Beggs and Brill gives good approximation with the ﬁeld data 12 in temperature. a better estimation of temperature proﬁle is obtained with the Fontanilla and Aziz approach. Heat transfer equations are developed by making an analogy with circuits for oﬀshore application and simple heat loss calculation is done with insulated and uninsulated tubing. The lower the thermal conductivity. The Fontanilla and Aziz approach is augmented and tested against the two ﬁeld cases. pressure calculations obtained over predicted steam quality results based on Fontanilla’s approach. Pressure and temperature distribution values converged to the ﬁeld data 1 and 2. more published ﬁeld data is needed. The other twophase correlations give similar values with Fontanilla’s results in terms of comparing steam quality. with using the Beggs and Brill model. CONCLUSIONS 103 increases. In order to see this. 7. Therefore.2 Conclusions 1. and we lose more heat to the surroundings.7. the lesser the heat loss to the surroundings. steam quality drops. in order to get more accurate simulation results. Uninsulated tubing gives 8 times higher heat loss when compared to the worst insulation material and 30 times more heat loss with compared to the best insulation material. Four diﬀerent twophase ﬂow correlations yield almost same heat loss . 2. and 600o F and plotted the results. In our calculations. 500o F . However.2. With improved twophase ﬂow correlations. we used three diﬀerent wellhead temperatures of 400o F . Heat loss and steam quality are highly related with the steam temperature and steam pressure. Beggs and Brill overpredicts steam quality values with compared to Fontanilla’s results. As we know each well has unique characteristics and production or completions history that makes predictions a little bit harder.
Using insulation materials only aﬀected heat losses and steam quality for both onshore and oﬀshore environments using insulated or uninsulated tubing.104 CHAPTER 7. Therefore. 6. SUMMARY. Identical wellhead steam injection temperature values for oﬀshore insulated tubing are considered with addition of 10 % N2 to steam and a pure steam only case. Field data of steam quality values are not available in the literature.37. 5. Formation pressure is important for both onshore and oﬀshore steam injection operations. 7. Onshore and oﬀshore results for both insulated or uninsulated cases showed similar trends in steam quality. 3. N2 . decreases saturated . CONCLUSIONS AND FUTURE WORK results in both ﬁeld cases. Having low thermal conductive material such as black aerogel provides signiﬁcantly less heat loss and steam quality decrease over the length of the well. The Beggs and Brill model yields higher steam quality values. steam pressure. and see the other parameters eﬀect on steam injection operations. One case is studied for two diﬀerent environments. The use of noncondensable gas. Sensitivity analysis of the diﬀerent parameters are conducted. We observed that having 10 % N2 increases steam quality value around 8 %. is added to steam is explored. steam temperature and heat loss proﬁles. and it is shown that when the formation pressure is greater than steam injection pressure at given depth steam can not be delivered to the reservoir. The calculated steam quality values based on all the investigated twophase ﬂow correlations are compared with the results of Fontanilla and Aziz. A GUI is developed for users to make the calculation faster. formation pressure importance is emphasized in Figure 6. while all the other models yield similar results with the base Fontanilla and Aziz calculation. 4.
At the same time. more ﬁeld data has to be provided by the industry. in order to ﬁnd best combination with regarding cost constraints and technology. 7.7. Changing depth do not have any contribution on both insulated tubing and uninsulated tubing cases. The trend with diﬀerent noncondensable gas mole fraction. and decreases heat loss around 2 %. rate. because the temperature diﬀerence between the well and the formation is greater. FUTURE WORK 105 steam pressure around 20 %. Sensitivity analysis and optimization can be done with diﬀerent injection temperature. With this more accurate results can be obtained. in order to expand of understanding of the physics of the problem. time. the greater the injection temperature the greater the pressure values and the smaller the change on steam pressure and steam temperature. diﬀerent injection rate.3. For uninsulated tubing cases. diﬀerent depth and diﬀerent injection temperature values are similar both insulated and uninsulated cases in for onshore and oﬀshore environments. Increasing steam temperature leads to greater heat loss. steam quality and depth. Greater the steam injection rate caused greater overall heat loss. . increasing N2 mole fraction does not change the bottomhole steam quality and yielded greater overall heat loss.3 Future Work For future work. The greatest eﬀect on bottomhole conditions of steam quality is observed with the highest wellhead steam temperature. diﬀerent steam quality.
dimensionless . dimensionless time conduction function also known as Ramey function twophase friction factor. o F/f t area. BT U/lb heat loss to the surrounding. CONCLUSIONS AND FUTURE WORK Nomenclature Abbreviations 2ResWithIns Sea ﬂoor to the reservoir with insulation 2ResWithoutIns ﬂoor to the reservoir without insulation Sea SeaWithIns Oﬀshore cases with insulation SeaWithoutIns Oﬀshore cases without insulation SLWithIns Surface line with insulation SLWithoutIns Surface line without insulation HSE VIT health.f t2 inclination factor noslip liquid holdup rate of heat loss per unit length of pipe in pressure change.106 CHAPTER 7. SUMMARY. psi heat loss to the surrounding. BT U/lb horizontal liquid holdup actual liquid holdup friction factor. safety and environment vacuum insulated tubing Symbols a A B(θ) CL ˙ Qls dp dQ dq EL (0) EL (θ) f f (tD) ftp geothermal gradient.
BT U/f t2 − hr radiation and convection coeﬃcient of heat transfer. f t outside radius of the casing.3. f t3 /hr inner radius of the tubing. BT U/f t2 − hr coeﬃcient of heat transfer forced convection. .7. f t wellbore radius. FUTURE WORK 107 fN S Frm g Gr h hf hf c hpipe hc.an HL J L k khc kha KE NRe P PE Pr qg rti rto rins rci rco rh no slip friction factor. f t outer radius of the tubing. BT U/lb Reynolds number pressure. BT U/(f t − hr −o F ) actual thermal conductivity of the annular ﬂuid. BT U/lb ﬁlm coeﬃcient of heat transfer of the pipe.17e − 8 f t/hr2 Grashoﬀ’s number enthalpy. f t . 778 f t − lbf /BT U length. 4. f t thermal conductivity of the material. BT U/f t2 − hr liquid holdup density mechanical equivalent of heat. BT U/f t2 − hr coeﬃcient of the heat transfer of pipe. psi potential energy. BT U/(f t − hr −o F ) kinetic energy. f t insulation radius of the tubing. f t inside radius of the casing. BT U/lb Prandtl’s number gas ﬂow rate. dimensionless Froud number of mixture acceleration due to gravity. BT U/(f t − hr −o F ) eﬀective thermal conductivity of the annular ﬂuid.
f t/hr superﬁcial velocity for gas phase. f t/hr superﬁcial velocity for liquid phase. f t3 /lb wind velocity. o F bulk temperature of the ﬂuid in the pipe.108 CHAPTER 7. f t/hr actual velocity for liquid phase. f t/hr actual velocity for gas phase. CONCLUSIONS AND FUTURE WORK rEa R Rh RN S t T Ta TA Tb Tm U v vw V vsg vg vsL vs vm Wm X z radius of the altered zone in the earth near the well. o F overall coeﬃcient of heat transfer. SUMMARY. BT U/hr − f t2 −o F speciﬁc volume. f t Greek Symbols αE thermal diﬀusivity of the earth. mph velocity. f t2 /hr blackbody emissive power. hrs temperature. f t Reynolds number speciﬁc thermal resistance noslip Reynolds number time. f t/hr mixture velocity. o F mean surface temperature. o R =o F + 460 ambient temperature of the atmosphere. o F absolute temperature. f raction by weight elevation or depth. f t/hr steam injection rate. BT U/hr − f t2 . lb/hr steam quality.
lb/f t − hr gas density.BT U/(f t − hr −o F ) gas viscosity. −90 λpipe λins λcem λEa λE µs µw ρl ρl ρs ρN S σ θ . BT U/hr − f t2 thermal conductivity of the pipe. 0. FUTURE WORK 109 ins ci emissivity of the insulation material .BT U/(f t − hr −o F ) thermal conductivity of the cement. lbm/f t3 slip mixture density.BT U/(f t − hr −o F ) thermal conductivity of the insulation. lbm/f t3 Stefan Boltzman constant. lbm/f t3 liquid density.7.BT U/(f t − hr −o F ) thermal conductivity of the unaltered zone. BT U/hr − f t2 emissivity of the casing .1714e − 8 BT U/(hr − f t2 −o R4 ) angle from horizontal.BT U/(f t − hr −o F ) thermal conductivity of the altered zone.lb/f t − hr liquid viscosity.3. lbm/f t3 noslip mixture density.
Appendix A Derivation of the Equations
A.1 Total Energy Equation
In Appendix A, all the equations are listed that are adapted and modiﬁed from the work that was done by Fontanilla and Aziz [19]. For a constant injection rate the continuity equation may be written as d 2 ρtp Vm rti = 0 dz (A.1.1)
where, ρtp is the twophase density (lb/ft3 ), Vm is the mixture velocity (ft/hr), and rti is the inside tubing radius (ft). The energy conservation equation may be derived by taking a diﬀerential element ∆z of the tubing as in Figure A.1: Energy In (BTU/lb) = Energy Out (BTU/lb)
2 z1 Vm1 z2 V 2 + = hm2 + + m2 J 2gJ J 2gJ
hm1 +
(A.1.2)
where, hm1 is the mixture enthalpy at point 1 (BTU/lb), z1 is the elevation at point 1 (ft), J is the mechanical equivalent of heat = 778 (ftlbf /BTU), Vm1 is the mixture 110
A.1. TOTAL ENERGY EQUATION
111
velocity at point 1 (ft/hr), g is the acceleration due to the gravity = 4.17*108 (ft/hr2 ), hm2 is the mixture enthalpy at point 2 (BTU/lb), z2 is the elevation at point 2 (ft), Vm2 is the mixture velocity at point 2 = Vm1 + dVm (ft/hr), dQ is the heat loss to the surroundings, BTU/lb.
Figure A.1: Schematic view of tubing element in our calculation.
2 Transferring the L.H.S. of Eq. A.1.2 to the R.H.S. and assuming dVm is negligible
we have, hm2 − hm1 + z2 − z1 Vm1 dVm + + dQ = 0 J gJ (A.1.3)
∆Enthalpy + ∆PE + ∆KE + Heat Loss to Surrounding = 0 The sign of the second term, ∆PE, is correct only if the elevation decreases downward. In the present problem for the seaﬂoor to reservoir, the surface is the datum plane and the elevation increases downward, so we need to aﬃx a negative sign before the
112
APPENDIX A. DERIVATION OF THE EQUATIONS
term. dhm + dividing by dz dhm 1 Vm1 dVm dQ + + + + =0 dz J gJ dz dz The kinetic energy term in Eq. A.1.5 may be expanded as : Vm1 = VSL + VSG = GL vL + Gg vg (A.1.6) (A.1.5) dz Vm1 dVm + + dQ = 0 J gJ (A.1.4)
where, VSL is the superﬁcial liquid velocity (ft/hr), VSG is the superﬁcial gas velocity (ft/hr), GL is the liquid mass ﬂux rate (lb/hrf t2 ), vL is the speciﬁc volume of liquid (f t3 /lb), Gg is the gas mass ﬂux rate (lb/hrf t2 ), vg is the speciﬁc volume of gas (f t3 /lb). Its equivalent diﬀerential form is dVm = GL dvL + Gg dvg Therefore we have Vm1 dVm 1 dVg dVL dVg dVL = + vL GL Gg + vg Gg GL + vg G2 vL G2 g L gJ dz gJ dz dz dz dz Substituting Eq. A.1.8 in to Eq. A.1.5 dVL dVg dhm 1 1 Vm1 dVg dVL dQ − + + + vL G2 + vL GL Gg + vg Gg GL + vg G2 + =0 L g dz J gJ gJ dz dz dz dz dz (A.1.9) (A.1.8) (A.1.7)
A.2. MECHANICAL ENERGY BALANCE OR THE EXTENDED BERNOULLI EQUATION113
A.2
Mechanical energy balance or the Extended Bernoulli Equation
The basis for any ﬂuid ﬂow calculation is the mechanical energy balance for the ﬂowing ﬂuid between two points. Because there is no external work done on or by the ﬂuid, a steadystate mechanical energy balance equation in diﬀerential form may be written for 1 lbm of ﬂuid as: 144 dp vm dvm + dWf = 0 + dz + 1 ρtp g (A.2.1)
where dp= P2 P1 is the change in pressure (lb/in2 ), ρtp is the twophase density (lb/ft3 ), dz=z2 z1 is the change in elevation (ft), vm1 is the mixture velocity at point 1 (ft/hr), dvm is the change in mixture velocity (ft/hr), g is the acceleration due to gravity = 4.17*108 (ft/hr2 ), dWf is the frictional drag (ft). We again have to aﬃx a negative sign on the second term because the elevation increases downward. Multiplying Eq. A.2.1 by 144
ρtp , dz
we have (A.2.2)
ρtp vm1 dvm dWf dp − ρtp + + ρtp =0 dz g dz dz
dp dp dp dp or 144 dz  dz elevation of PE+ dz acceleration or KE+ dz friction = 0
Substituting Eq. A.1.9 into Eq. A.2.2, we have
144
dVL dVg dp ρtp dVg dVL dP −ρtp + vL G2 + vL GL Gg + vg Gg GL + vg G2 + L g dz g dz dz dz dz dz
f riction
(A.2.3) The mixture enthalpy hm is a function of steam quality X and pressure P (psia). hm = f (X, P ) (A.2.4)
114
APPENDIX A. DERIVATION OF THE EQUATIONS
The speciﬁc volume of liquid (water) and gas (steam) are just functions of pressure P (psia). VL = g (P),Vg = h (P) thus we can write ∂hm dX ∂hm dP dhm = + dz ∂X dz ∂P dz dvg ∂vg dP = dz ∂P dz dvL ∂vL dP = dz ∂P dz Substituting Eq. A.2.5, Eq. A.2.6, and Eq. A.2.7 into Eq. A.1.9 yields: ∂Vg ∂VL ∂hm dX ∂hm dP 1 1 dP ∂Vg ∂VL dQ − − + + vL GL Gg + vg Gg GL + vg G2 =0 vL G2 + g L ∂X dz ∂P dz J gJ dz ∂P ∂P ∂P ∂z dz (A.2.8) Substituting Eq. A.2.6, and Eq. A.2.7 into Eq. A.2.3 we have: 144 dP ρtp dP ∂VL ∂Vg ∂VL ∂Vg dP =0 −ρtp + vL G2 + vL GL Gg + vg Gg GL + vg G2 + L g dz g dz ∂P ∂P ∂P ∂P dz f riction (A.2.9)
dP dz
(A.2.5)
(A.2.6) (A.2.7)
Solving for
in Eq. A.2.9, ρtp −
ρtp g dP dz f riction
dP = dz 144 + Solving for dX dz = 1
dhm dX dX dz
vL G2 ∂VL + vL GL Gg ∂Vg + vg Gg GL ∂VL + vg G2 ∂Vg g ∂P L ∂P ∂P ∂P
(A.2.10)
in Eq. A.2.8 yields, vL G2 L ∂VL ∂Vg ∂VL ∂Vg + vL GL Gg + vg Gg GL + vg G2 g ∂P ∂P ∂P ∂P (A.2.11)
−
∂hm dP 1 1 dP + − ∂P dz J gJ dz
−
dQ dz
Eq. A.2.10 and Eq. A.2.11 are two simultaneous, ﬁrst order ordinary diﬀerential equations to be solved for pressure (P) and quality (X). Besides these we need to ﬁnd
2πke (Th − Te ) dq = z f (t) (A.1 refers to surface lines.3. the rate of heat conduction into the earth is expressed as. In the present case. dq is the change in heat loss rate (BTU/hr). Figure 3. A. we need to know to the surroundings. ke is the thermal conductivity of the .3. dQ = (BTU/lbft) which is the heat loss dq Wm (A. dz A. A.A. EVALUATION OF HEAT LOSS TO THE SURROUNDING 115 a method to calculate dQ .4 refers to sea level to sea ﬂoor and Figure 3.11.2 is based on any reference point.3 Evaluation of Heat Loss to the Surrounding dQ dz In order to solve Eq. Wm is the steam injection rate (lb/hr) Because we have a pseudo steady state heat ﬂow conditions in the wellbore.2. From Ramey[42].3.3. Figure 3.3. The radius r and overall heat transfer coeﬃcient U used in Eq. the outer tubing surface was chosen as the reference plane.9 refers to sea ﬂoor to reservoir heat loss calculation with insulated or not.1) dQ is the change in heat loss by ﬂuid (BTU/lb). We adapted the surface line heat losses from Prats and the rest of theprocedures are based on Willhite’s approach that was discussed in chapter 3.2) A representation of the problem for three parts of our system will be analyzed and we refer those ﬁgures here again.3) where: Te is the temperature of the earth ( o F). Three ﬁgures and having two options each of them gives six diﬀerent heat loss calculation procedures. the rate of heat conduction from the ﬂuid to cementformation interface (hole) is expressed: dq = 2πrto (Tf − Th ) dz (A.
3.6). f(t) is the time conduction function.6) A. The rate of the heat transfer between the ﬂowing ﬂuid and the inside tubing is given as: dq = 2πrti hf (Tf − Ti ) dz (A.3. The rate of heat conductance from the ﬂuid to the cementformation (hole) interface is given by A.3.3.4.2 and Eq. Tm is the mean surface temperature (o F). A. Because the rate of heat conduction from the ﬂuid to the cement formation interface (hole) must equal the rate of heat conduction into the earth. is given by Te =Tm +az. A.2 we have: 2πrto Uto dQ rto Uto f (t)Tf + ke Te = Tf dz Wm rto Uto f (t) + ke (A.3. the temperature of the earth.3.5 we have: 2πrto Uto rto Uto f (t)Tf + ke (Tm + az) dQ = Tf dz Wm rto Uto f (t) + ke (A.3. A. otherwise it assumes constant value. a is the geothermal gradient ((o F)/ft). The overall heat transfer conductance Uto may be evaluated by writing expressions for the heat ﬂow through the various components of the injector.3 may be equated to obtain an expression for Th : Th = rto Uto f (t)Tf + ke Te rto Uto f (t) + ke (A.4 Determination of the Uto and Tci dQ dz To obtain (Eq. A.2.116 APPENDIX A.3.3.5) If the geothermal gradient is provided.4) Substituting Eq.3. we need to know Uto which is the overall heat transfer coeﬃcient or the overall coeﬃcient of conductance. Te .3. Eq. z is the depth (ft). A.4 and Eq.1 into Eq.1) . A. DERIVATION OF THE EQUATIONS earth (BTU/hrfto F). A. When we substitute the temperature of the earth into Eq.
On the other hand. heat conduction through the ﬂuid in the annular space. casing wall. insulation. and the cement sheath occurs by conduction. Heat ﬂow through the tubing wall.4.A.3) . radiation Heat transfer by natural convection in the annulus is caused by ﬂuid motion resulting from the variation of density with temperature. DETERMINATION OF THE UT O AND TCI 117 where hf is the ﬁlm coeﬃcient for heat transfer based on the inside surface area of the tubing. 2. When a body is heated. natural convection of ﬂuid 3.4.2) Integrating Fourier’s law of heat conduction over the thickness of the insulation yields: dq 2πkins = (Tto − Tins ) dz ln( rrins ) to There are three modes of heat transfer present in the casing annulus. The amount of radiant energy transported between the tubing/insulation and casing depends on the view the surfaces have of each other and their emitting and absorbing characteristics. the ﬂuid near the casing wall is cooler and denser than that in the center of the annulus and tends to fall. To account for these three modes of heat transfer. 1. radiant energy is emitted at a rate dependent on the temperature of the body.4. Integrating Fourier’s law of heat conduction over the thickness of the tubing wall we write: 2πktub dq = (Tti − Tto ) dz ln( rto ) rti (A. Hot ﬂuid near the tubing and insulation is less dense than the ﬂuid in the center of the annulus and tends to rise. it is convenient to deﬁne the heat rate through the annulus in terms of heat transfer coeﬃcients hc (natural convection and (A.
11) Tins − Tci = 2πrto (hc + hr ) (A.5) (A.8) Tf − Tti = Tti − Tto = Tto − Tins = 2πrti hf ln rto rti rins rto (A.6) 2πrto Uto dq dz (A.4) (A.4.4. DERIVATION OF THE EQUATIONS conduction) and hr (radiation).4.4. as: dq = 2πrto (hc + hr )(Tins − Tci ) dz The heat conduction through the casing wall is given as: dq 2πkcas = (Tci − Tco ) dz ln( rco ) rci The heat conduction throught the cement sheath is given as: 2πkcem dq = rh (Tco − Th ) dz ln( rco ) Note that: Tf −Th = (Tf −Tti )+(Tti −Tto )+(Tto −Tins )+(Tins −Tci )+(Tci −Tco )+(Tco −Th ) (A.4.118 APPENDIX A.7) Solving for the temperature diﬀerences Tf − Th = dq dz (A.4.9) dq dz 2πktub dq dz (A.12) .4.4.10) 2πkins ln dq dz (A.4.
2 to 0.7 we will have: 1 + 2π rti hf dq dz ln Tf − Th = rto rti ln + rins rto ktub kins + 1 + rto (hc + hr ) ln rco rci ln + rh rco kcas kcem (A.15) An equation for Uto is found by comparing Eq.02 to 0.8 and Eq.4.6 BTU/hrfto F ).17) The thermal conductivity of the tubing and casing steel (ktub = kcas = 25BT U/hr − f t −o F ) is considerably higher than that for the other materials in the wellbore (insulation = 0.A. A.4.9 . A.4.4.4. A.4.4.4. A.13) 2πkcem ln (A. . Therefore. DETERMINATION OF THE UT O AND TCI 119 Tci − Tco = Tco − Th = dq dz 2πkcas dq dz ln rco rci rh rco (A. Uto = rto + rti hf rto ln rto rti rto ln + rins rto ktub kins + 1 + (hc + hr ) rto ln rco rci rto ln + rh rco −1 kcas kcem (A.4.16) Therefore.4.15 1 1 = + rto Uto rti hf ln rto rti ln + rins rto ktub kins + 1 + rto (hc + hr ) ln rco rci ln + rh rco kcas kcem (A. its relative contribution in Eq.14 into the Eq.17 is small and can be neglected (Tti = Tto and Tci = Tco ).14) Substituting Eqs.06 BTU/hrfto F and cement = 0.4.4. A.
Thus Eq.5.2 may also be expressed if there is no insulation as: dqc = 2πrto hc (Tto − Tci ) dz if there is a insulation: dqc = 2πrins hc (Tins − Tci ) dz (A. A.18.5 Determination of the Convection Heat Transfer Coeﬃcient Before Uto can be calculated in Eq. Heat transfer per unit length by conduction and free convection in the annulus is given if there is no insulation.4.4.5. 2πkhc dqc = rci dz ln rto if there is an insulation.18 Uto = rto ln rins rto kins + 1 + (hc + hr ) rto ln rh rco −1 (A.4) (A.3) . A. A.18) kcem A.1) We can also express Eq.2) (A.5. A. dqc 2πkhc (Tins − Tci ) = rci dz ln rto (A.120 APPENDIX A.1 and Eq.5.17 simpliﬁes to Eq.4.5. A.5. the convection coeﬃcient hc and the radiation hr must be evaluated. DERIVATION OF THE EQUATIONS Also the ﬁlm coeﬃcient hf for steam and water are high enough (500 to 4000 BTU/hrf t2 o F ) to justify the assumption of inﬁnite ﬁlm coeﬃcient (Tf = Tti ).4.
5. β= 1 (Tan + 460) (A.5.4 we will have if there is no insulation. β (o R−1 ) is the coeﬃcient of volume expansion. hc = if insulation present. For ideal gas.333 (P r)0.8) (A.5. A.2 with Eq.5.049)(GrP r)0. A. DETERMINATION OF THE CONVECTION HEAT TRANSFER COEFFICIENT121 Combining Eq.5.10) (rci − rins )3 gρ2 β(Tins − Tci ) an µ2 an (A. A.074 where: Gr is the Grashoﬀ’s number.5.5.7) where Can is the speciﬁc heat of the annular ﬂuid and µan is the viscosity of the annular ﬂuid with given pressure in the annulus.5.5.5) (A.5.3 and Eq. khc = (kha)(0.6) The eﬀective thermal conductivity of the annular ﬂuid (khc) is related to the actual thermal conductivity of the annular ﬂuid (kha) as a function of the Grashoﬀ number and Prandtl number.5.5.1 with Eq.9) (rci − rto )3 gρ2 β(Tto − Tci ) an µ2 an (A. if there is no insulation Gr = if there is an insulation Gr = Pr is the Prandtl’s number Pr = Can µan kha (A.A. A. β is the reciprocal average absolute annulus temperature. hc = khc rins ln rci rins khc rto ln rci rto (A.11) .
6.2) −1 A. DERIVATION OF THE EQUATIONS where: Tan if there is no insulation Tan = if there is an insulation Tan = Tins + Tci 2 (A.6. hr and hence Uto .. Since we need to .7 Computational Procedure for Uto In order to calculate hc .5. The following developments are only for bare tubings but an equivalent equation which can be used for calculation of hr for tubings with insulation is also given. So hr without insulation hr = (T ato + T aci )(T a2 + T a2 )σ to ci 1 to + rto rci 1 ci (A.12) A.122 APPENDIX A.5. et al.1) −1 hr with insulation hr = (T ains + T aci )(T a2 + T a2 )σ ins ci 1 ins + rins rci 1 ci (A. Here we are going to give the equations for hr with insulation and without insulation.13) Tto + Tci 2 (A.6 Determination of the Radiation Heat Transfer Coeﬃcient Radiation heat transfer problems can be represented by ”radiation networks” as in Carslaw and Jaeger and Herrera. we need to know the temperature of the inside casing Tci and the temperature at the outer insulation Tins . The tubing and the casing surfaces which exchange heat with each other may be modelled by resistances in series as shown in Fontanilla’s thesis [19].
3.7. COMPUTATIONAL PROCEDURE FOR UT O 123 know Tci and Tins (which are unknown) to solve for hc . Before describing the iteration procedure.4. (A.2) Substituting this expression for Tco in Eq.7. Tci and Tins . (A. a trial and error or an iterative solution is required to determine the proper combination of Uto .4.7. the equations to be used the iteration are developed.7.3) in Eq.7.4) By negleccting the thermal resistance of the ﬁlm.6): dq ln dz rh rco Tins = Th + 2πkcem (A. (A. Eq. (A. we have: dq ln dz rco rci dq ln dz rh rco Tci = Th + Substituting Eq.7.5) The hole temperature (Th ) can be obtained from Eq.2) for dq dz 2πkcas + 2πkcem (A.7. and casing. The equation for Tins is obtained from Eq.A.6) . Th = rto Uto f (t)Tf + ke Te rto Uto f (t) + ke (A. hr and Uto .7. and solving for Tci . tubing.4) derived earlier.7. (A. (A.3) and solving for Tci ln rco rci kcas + rh ln rco Tci = Th + rto Uto (Tf − Th ) + kcem (A.4) reduces to : Tci = Th + rh rto Uto ln rco kcem (Tto − Th ) (A. (A.5).3.3): dq ln dz rins rto Tins = Tto − 2πkins (A.4.1) The equation for Tco is obtained from Eq.
if newTci . (A. (A.18) and f(t) from Eq.124 APPENDIX A.5). In summary.oldTci  > 1. Tins and Tci .1). 2. (A. Having values for Tins and Tci .5.7.2).3. hc . (A. If the absolute value of their diﬀerence is greater than a tolerable amount.6) Uto from Eq. Calculate the old Th from Eq. the iteration procedure is as follows: 1. Calculate hr from Eq. Assume an arbitrary value for old dq dz and calculate the old Tins from Eq.6.7. DERIVATION OF THE EQUATIONS The ﬁrst thing to do is to assume an arbitrary value for dq dz and calculate an initial Tins from Eq. solving for Th in Eq.7) Using the new Th from Eq. (A. Th is then solved using Eq. The old Tci and new Tci are compared.8. (A.7. (A.3.2). The old and new casing temperatures are compared at this point. (A.7. Then set Tci equal to the geothermal temperature (Te ). hc from Eq. The iteration is continued until convergence is obtained.6).7. Uto .7. and f(t) can be calculated.1) knowing the old Tins and the old Tci 4. .5). Otherwise convergence is obtained.7). (A. To obtain a corresponding Th based on the new Tci .7. 6. (A. Then a new value of Tci is calcualted using Eq. (A. we calculate from Eq. hr .7.4) 5. Calculate the dq new Tins from Eq.7.1) using the new dz . the old Tci is incremented by 70 percent of the absolute value of the diﬀerence. Calculate the new Tci using Eq. The model then uses this new casing temperature.1).6) yields: Th = rh Tci kcem − rto ln rco Tto rh kcem − rto Uto ln rco dq dz (A. say 1 o F. the old casing temperature is incremented by 70 percent of the diﬀerence. (A.4. (A. Set Tci equal to the geothermal temperature Te and call this the old Tci 3. (A.
A.8.8.1) 10. Starting with step 3. Calculate Tins from Eq. the procedure is repeated until convergence is obtained. 9. the convergence time is of the order of one week for many reservoir problems.3. A.7. a corresponding Th based on the Tci incremented in step 6. (A. is calculated using Eq. From Ramey[42] an approximate equation for f (t) satisfying the line source solution for long times is: f (t) = −ln r √h 2 αt (A. Calculate a corresponding dq dz from Eq.1) . (A.8 Determination of f(t) dQ dz The time conduction function f(t) introduced in the equation of unsteady state heat ﬂow to the earth and needed to obtain and Uto can be estimated from solutions for radial heat conduction from an inﬁnitely long cylinder. (A. As can be seen Figure A. Such solutions are analogous to transient ﬂuid ﬂow solutions used in reservoir engineering. 8. If convergence is not yet obtained. DETERMINATION OF F(T) 125 7. Thus. the line source solution will often provide useful results if times are greater than a week.7).5 all ﬁve solutions converge to the same line.2).7.
5. In this model.2: Time conduction function (retrieved from[19]). regression was used to obtain a third order polynomial approximation for each of the ﬁve curves in Figure A.126 APPENDIX A. DERIVATION OF THE EQUATIONS Figure A. Lets say K = rh U ke the calculation procedure is as follows: .
19865 + . EVALUATION OF THE DERRIVATIVES 127 Algorithm 1 f (t) calculation based on the regression analysis Require: x = log10 αt 2 rh and Y = log10 f (t) Ensure: f (t) = 10Y 1: if K == 0 then 2: Y ← .33116x − .033723x2 − .48034x − .000019P −0.06586x2 − .9.07525x2 + .9 Evaluation of the Derrivatives Farouq Ali[19] presented approximate correlations for speciﬁc volumes of saturated liquid and saturated vapor.4) .45 vL = 363.12557 + .0 then 14: Y ← −.9.3) (A. ∂vL = 0.000086P 0.907P −1.A.00222x3 9: end if 10: if K ≤ 5.38757x − .00148x3 3: else 4: if K ≤ .3689x − .36166x − .0002P 0.9.2) (A.0009P −0.9.00393x3 12: end if 13: if K ≥ 5.225 + 0. vL = 0.00525x3 15: end if 16: end if A.03018 + .1) (A.08619x2 + .9.0 then 11: Y ← −.775 + 0.55 ∂P ∂vg = −348.01065x3 6: end if 7: if K ≤ 1.0 then 8: Y ← −.08738 + .9588 Diﬀerentiating with respect to pressure we have.9588 ∂P (A.01587 + 0.2 then 5: Y ← −.9P −0.04619x2 − .02435 + .
9. we have: ∂hm = 1318P −0.5) (A.6) (A.10 Calculation of the Annulus Fluid Properties There are several modiﬁcation has been done for Fontanilla[19] solutions. We get viscosity and thermal conductivity of the ﬂuid under 1 atm using ﬁgures from Prats[41].9.7426 − 115.08774 ) Diﬀerentiating with respect to the steam quality X and pressure P.08774 ∂P (A. DERIVATION OF THE EQUATIONS Farouq Ali[19] also proposed correlations for the enthalpy of saturated liquid and the heat of vaporization Lv hL = 91P . .423P −0.9.7) (A.128 APPENDIX A.08774 ∂X ∂hm = 23.8) (A.2574 + X(1318P −.2574 Lv = hg − hl = 1318P −.9.9) (A.64XP −1.08774 for wet steam.9. hm = hL + XLv hm = 91P . one of them is getting annulus ﬂuid properties automatically.9.10) A.
0170 µ1 − µ2 = = 2.0184 − 0.1) .0192 cp.8 ∗ 10−5 T1 − T2 100 − 50 (A.A. and viscosity values are µ1 2 = 0. We took two points on our curve and showed on the ﬁgure both N2 and air and assumed that the line is linear with increasing temperature values. For air same procedure N N applies µ1 =0.0170 cp and µ2 2 = 0.0176 cp and µ2 =0. CALCULATION OF THE ANNULUS FLUID PROPERTIES 129 Viscosity of the annular ﬂuid. Temperature values both N2 and air are T1 = 50 and T2 = 100. µannulus Figure A.3: Viscosity of the annular ﬂuid with respect to Temperature. Let’s ﬁnd mN2 and mair values as follows: air air mN2 = 0.10.0184 cp.10. As we know from very basic way to get slope of the line and put that value into y = mx + n equation to get the correlation between temperature. because lines are only slightly diﬀerent from linearity.
5 N BTU/fto F D and λ2 2 = 0.68 BTU/fto F D. and thermal conductivity values are λ1 2 = 0.2 ∗ 10−5 T1 − T2 100 − 50 (A.2) Thus. We used this correaltions wrt temperature in our calculation. Temperature values both N2 and air are T1 = 400 and T2 = 800. λannulus Figure A.5 air N .0176 µ1 − µ2 = = 3.2 ∗ 10−5 T+0. The same procedure is also applied for getting correlation both N2 and air for thermal conductivity with respect to temperature.8 ∗ 10−5 and the correlation we got µN2 =2.10.0192 − 0. For air same procedure applies λ1 =0. DERIVATION OF THE EQUATIONS mair = 0. slope mN2 = 2.8 ∗ 10−5 T+0.130 APPENDIX A. Thermal conductivity of the annular ﬂuid.0176.4: Thermal conductivity of the annular ﬂuid with respect to Temperature.0170 and for air µair =3.
CALCULATION OF THE ANNULUS FLUID PROPERTIES 131 BTU/fto F D and λ2 =0.71 − 0.4) Thus.71 BTU/fto F D.10.5 = = 5.10.10. Let’s ﬁnd mN2 and mair values as follows: air mN2 = 0. .5 ∗ 10−4 T1 − T2 800 − 400 λ1 − λ2 0.50 λ1 − λ2 = = 4.5 ∗ 10−4 T+0.25 ∗ 10−4 T+0. slope mN2 = 4.A.68 − 0.25 ∗ 10−4 T1 − T2 800 − 400 (A.5 ∗ 10−4 and the correlation we got λN2 =4.5 and for air λair =5.5.3) mair = (A. We used this correaltions wrt temperature in our calculation.
Appendix B Codes for Heat Loss Calculations B.1: Schematic representation of resistance to heat transfer with/without temperature proﬁle. (b) Surface pipe without insulation. Figure B.1 Heat Losses from Surface Line (a) Surface pipe with insulation. 132 .
4 hfcWins = ( 1 8 ∗ ( AvrWindSpeedWins ˆ 0 . = 600/24. % We a r e l o o k i n g f o r t h e s o l u t i o n a t s t e a d y s t a t e c o n d i t i o n s % f o r pipe i s insulated % Thermal Recovery by Micheal P r a t s c h a p t e r 1 0 . % h f c c a l c u l a t i o n based on t h e Eq .166 0.4167. close a l l .1. = 2000. . 0. 6 ) ∗ ( r i n s u l a t e d W i n s ˆ . = inf . f o r i =1: length ( lambdaInsWins ) % Overall s p e c i f i c thermal r e s i s t a n c e .388 0. / 2 4 .6)/ rinsulatedWins )/24.569 0 . = 100.125 −136 % % clc . = 0. TsteamWins TaverageWins PipeLengthWins InjTimeWins = 550. c o l o r s = l i n e s ( length ( lambdaInsWins ) ) . = 0. clear a l l . 1 C a l c u l a t i o n s o f Heat L o s s e s from S u r f a c e L i n e s % % 4 i n c h e s N−80 p i p e a t r a t e o f 229 B/D.B.499 0. % F % F % ft % hours % mph % f t from Table B. % Btu/ sq f t −hr−F % Btu/ sq f t −hr−F % Btu/ sq f t −hr−F CollectDatasWithins = [ ] . 1 5 % ft % ft % Btu/ f t −hr−F AvrWindSpeedWins = 2 0 .1478. = 2000. . rinsideWins routsideWins rinsulatedWins lambdaPipeWins lambdaInsWins hfWins hpiWins hpoWins = 0. 9 6 ] .194 0. = 365∗24. HEAT LOSSES FROM SURFACE LINE 133 The following are the MatLab codes for our heat loss calculation starting from surface lines with insulation and without insulation % Example 1 0 .1667. 1 0 . = [0. pgs . = 60.
end =(1: InjTimeWins ) ’ . 1 6 ) LambdaValStr = s p r i n t f ( ’ %0. ’ Fontname ’ . 1 / ( hfcWins ∗ r i n s u l a t e d W i n s ) ) .3G ’ .m: . . . . % Heat L o s s e s QlsWins = ( TsteamWins − TaverageWins ) / RhWins . InjTimeplotWins . l a b e l s { i } = [ ’ \ lambda = ’ num2str ( LambdaValStr ) ] . : ) .134 APPENDIX B. . ’ n o r m a l i z e d ’ ) . ’ U n i t s ’ . . . ’ FontSize ’ . ’ Fontname ’ . set ( gcf . lambdaInsWins ( i ) ) . % Amount o f Heat Lost from t h e p i p e o v e r a p e r i o d o f time QlWins = QlsWins ∗ InjTimeWins ∗ PipeLengthWins . ’ L o c a t i o n ’ . . . 2 . . ’ YDir ’ . legend ( l a b e l s . . [ RhWins QlsWins QlWins ] ] . ’ O u t e r P o s i t i o n ’ . . ’ c o l o r ’ . ’ L i ne w i dt h ’ . log ( r o u t s i d e W i n s / r i n s i d e W i n s ) + 1 / ( hpoWins∗ r o u t s i d e W i n s ) + . InjTimeplotWins PlotQlWins figure (1) plot ( PlotQlWins . 3 ) hold on grid on set ( gca . . num2str ( PipeLengthWins ) ’ f t ’ ] . . . CODES FOR HEAT LOSS CALCULATIONS RhWins = 1 / ( 2 ∗ pi ) ∗ ( 1 / ( hfWins ∗ r i n s i d e W i n s ) . + 1 / ( hpiWins ∗ r i n s i d e W i n s ) + ( 1 / lambdaPipeWins ) ∗ . ’ C a l i b r i ’ . ’ top ’ . ( 1 / lambdaInsWins ( i ) ) ∗ log ( r i n s u l a t e d W i n s / r o u t s i d e W i n s ) + . ’ XAxisLocation ’ . C o l l e c t D a t a s W i t h i n s =[ C o l l e c t D a t a s W i t h i n s . c o l o r s ( i . . ∗ InjTimeplotWins . Code block: SLwithoutIns. . ’ r e v ’ ) xlabel ( [ ’ Cumulative Heat L o s s e s ( i n BTU) f o r ’ . . ’NE ’ ) set ( gcf . =(QlsWins ∗ PipeLengthWins ) . . ’ F o n t S i z e ’ . [ 0 0 1 1 ] ) . 16) ylabel ( [ ’ Time ( i n Days ) ’ num2str ( InjTimeWins / 2 4 ) ] . / 2 4 . ’ C a l i b r i ’ . .
= 0. % F % F % ft % days % mph % f t from Table B. TsteamWOutIns TaverageWOutIns PipeLengthWOutIns InjTimeWOutIns = 550. rinsideWOutIns routsideWOutIns r i ns u l a t e dW O u t I ns lambdaPipeWOutIns lambdaInsWOutIns hfWOutIns hpiWOutIns hpoWOutIns = 0. % % clc . . 1 C a l c u l a t i o n s o f Heat L o s s e s from S u r f a c e L i n e s % % 4 i n c h e s N−80 p i p e a t r a t e o f 229 B/D. . m o d i f i e d d a t e 2 0 1 1 . = 365. + 1 / ( hpiWOutIns ∗ rinsideWOutIns ) + ( 1 / lambdaPipeWOutIns ) .4167. . .96. = 0. clear a l l . = 100. ∗ ( routsideWOutIns ˆ 0 . HEAT LOSSES FROM SURFACE LINE 135 % Example 1 0 . 1 0 . 1 5 % ft % ft % Btu/ f t −D−F % Btu/ f t −D−F % Btu/ sq f t −D−F % Btu/ sq f t −D−F % Btu/ sq f t −D−F AvrWindSpeedWOutIns = 2 0 . = 48000. = inf . . = 0.1667. % We a r e l o o k i n g f o r t h e s o l u t i o n a t s t e a d y s t a t e c o n d i t i o n s % f o r p i p e i s not−i n s u l a t e d % w r i t t e n 2 0 0 9 . % Overall s p e c i f i c thermal r e s i s t a n c e RhWOutIns = 1 / ( 2 ∗ pi ) ∗ ( 1 / ( hfWOutIns∗ rinsideWOutIns ) . . = 48000.B.1.1478. May 9 th % Fidan . % h f c c a l c u l a t i o n based on t h e Eq . . = 600. htTable14 ( TsteamWOutIns . . S . . . close a l l . . 4 hfcWOutIns = 1 8 ∗ ( AvrWindSpeedWOutIns ˆ 0 . = 60. 6 ) / routsideWOutIns . ∗ log ( routsideWOutIns / rinsideWOutIns ) + 1 / ( ( hfcWOutIns + . . 6 ) . routsideWOutIns ) ) ∗ .
’ num2str ( RhWOutIns ) ’ (BTU/ f t −D−ˆF)ˆ−1 ’ ] ) % Heat L o s s e s QlsWOutIns /RhWOutIns . . % Amount o f Heat Lost from t h e p i p e o v e r a p e r i o d o f time QlWOutIns = QlsWOutIns∗ InjTimeWOutIns . QlWOutIns . d i s p l a y ( [ ’ Cumulative Heat L o s s e s o v e r t h e p e r i o d . 1 0 . . ’ Btu/ f t −D ’ ] ) = ( TsteamWOutIns − TaverageWOutIns ) . . . . CODES FOR HEAT LOSS CALCULATIONS routsideWOutIns ) ) . .136 APPENDIX B. set ( gcf . . . set ( gcf . . PipeLengthWOutIns . . . r e s i s t a n c e i s c a l c u l a t e d from Eq . ∗ InjTimeplotWOutIns . . . [ 0 0 1 1 ] ) . p l o t H e a t L o s s ( PlotQlWOutIns . d i s p l a y ( [ ’ The o v e r a l l s p e c i f i c t h e r m a l . ’ n o r m a l i z e d ’ ) . ’ U n i t s ’ . 2 i s = . ’ num2str ( QlWOutIns ) ’ Btu ’ ] ) h = figure ( ’ Color ’ . . . ’ r ’ ) . =(QlsWOutIns∗ PipeLengthWOutIns ) . o f time woth g i v e n p i p e l e n g t h Ql = . InjTimeplotWOutIns PlotQlWOutIns =(1: InjTimeWOutIns ) ’ . [ 0 0 0 ] ) . d i s p l a y ( [ ’ Heat L o s s e s Qls = ’ num2str ( QlsWOutIns ) . . . InjTimeplotWOutIns . . ∗ PipeLengthWOutIns . h . . . . lambdaInsWOutIns . ’ O u t e r P o s i t i o n ’ .
close a l l . r r i s e r o =18/12/2. r c i= r r i s e r i . HEAT LOSSES FROM SEA PART 137 B. May 9 th % Fidan . r r i s e r i =0. Figure B.2 Heat Losses from Sea Part (a) Sea Part with insulation. . rto = 0. (b) Sea Part without insulation.60. clear a l l . S . m o d i f i e d d a t e 2 0 1 1 .B.2: Schematic representation of resistance to heat transfer with temperature proﬁle. .2. % % clc .1458. The following are the MatLab codes for our heat loss calculation starting from sea surface to sea ﬂoor with insulation and without insulation % S o l u t i o n f o r O f f s h o r e p a r t with i n s u l a t i o n m a t e r i l a s % % We a r e l o o k i n g f o r t h e s o l u t i o n a t s t e a d y s t a t e c o n d i t i o n s % w r i t t e n 2 0 0 9 .
% c o n v e r s i o n f a c t o r LamdaTub=600/24. lamda EWithIns = ke . %[ Perry ’ s Chemical E n g i n e e r s ’ Handbook .499 0. r c o W i t h I n s= r r i s e r o . r c i W i t h I n s= r r i s e r i . % t h e r m a l c o n d u c t i v i t y o f water %i n [W/m K] ==1 Btu / ( h r f t ?F) = 1 . a l p h a 1 = S W D i f f u s i v i t y (T. / 2 4 . 9 . 9 6 ] . = 21∗24. 8 . = 164.138 APPENDIX B. 9 . rEaWithIns = 0 .388 0.2292. rinsWithIns = 0. CODES FOR HEAT LOSS CALCULATIONS rco = r r i s e r o . % hrs % F % ft .166 0. S=35.% i n [mˆ2/ s ] % 1 [mˆ2/ s ] =38750. roWithIns= r t o .569 0 . S ) .077500155 [ f t ˆ2/ hr ] from q1=a l p h a 1 ∗ 3 8 7 5 0 . 9 . rwWithIns = rh . 7 3 0 7 3 5 W/ ( mK ) . %[ f t ˆ2/ hr ] e p s c i = 0 . Table 1 −4] ke=k1 / 1 .194 0. S ) . r i n s=r i n s W i t h I n s . k1 = SW Conductivity (T. e c i W i t h I n s= 0 . e p s r i s e r = 0 . % Btu/ f t −hr−F = [0. T=(TAWithIns − 3 2 ) / 1 . = 600. 9 . TAWithIns = 7 0 . 9 . 0 7 7 5 0 0 1 5 5 . e p s t o= 0 . e i n s W i t h I n s= 0 . lamda insWithIns InjTimeWithIns TbWithIns PipeLengthWithIns k i n s=l a m d a i n s W i t h I n s . aEWithIns= q1 . 7 th E d i t i o n . rh= 1 . 7 3 0 7 3 5 .
%lbm/ f t ˆ3 Gr1=Gr ( T i n s o l d . kha ) . g = 3 2 . Beta Gr . TAWithIns .%cp % 1 cp = 1488 l b / f t −s e c o n d = 1488∗3600 l b / f t −hr visAn = visAn ∗ 1 4 8 8 ∗ 3 6 0 0 . . lamdaPipe = 6 0 0 / 2 4 . HEAT LOSSES FROM SEA PART 139 t=InjTimeWithIns . e p s c i . Boltzman ) . % c a l c u l a t e t h e o l d T i n s from eq . den=densityAn ( Tanulus ) . 1 7 1 4 e −8). den . TAWithIns ) .%R visAn=v i s c o s i t y A n ( Tanulus ) . Pr1 . 16 T i n s o l d=TAWithIns . CollectDatasInjIns =[]. Pr1=Pr ( kha . 9 Tanulus=Tan ( T i n s o l d . % hc from eq . % Tto = TbWithIns . r i n s . 2 ∗ 3 6 0 0 ∗ 3 6 0 0 . khc=khc1 ( Gr1 . Boltzman = ( 0 . Th=TAWithIns .B. visAn ) . % h r from eq . Tf = TbWithIns . 15 h r=hr ( T i n s o l d . TAWithIns . r c i . r c i . visAn . e p s t o . Tgeo=TAWithIns .%F Beta Gr=BetaGr ( Tanulus ) . kha=kha1 ( Tanulus ) . r i n s .2. f o r i =1: length ( l a m d a i n s W i t h I n s ) % Step 1 a s s i g n random v a l u e f o r dqdz dqdz =5. c o l o r s = l i n e s ( length ( l a m d a i n s W i t h I n s ) ) . g ) .
else while abs ( Tins new−T i n s o l d )>1 i t e r=i t e r +1. . . 8 Uto=UtoCal ( h c . Tins new=Tto −(( dqdz ∗ log ( r i n s / r t o ) ) / ( 2 ∗ pi ∗ k i n s ( i ) ) ) .%F Beta Gr=BetaGr ( Tanulus ) . r i n s ) . Boltzman ) . h r=hr ( T i n s o l d . % f ( t ) from Ramey and W i l l h i t e ftD = FTIME( Uto . dqdz=2∗pi ∗ r t o ∗Uto ∗ ( Tf−Th ) . Th=T i n s o l d −dqdz / ( 2 ∗ pi ∗ r t o ∗ ( h c+h r ))+ dqdz ∗ log ( r r i s e r o / . e p s c i . . k i n s ( i ) ) . % from eq . r r i s e r i ) / ( 2 ∗ pi ∗LamdaTub ) . CODES FOR HEAT LOSS CALCULATIONS h c=hc ( khc . . Tins new=Tins new1 . Tto=TAWithIns+dqdz / ( 2 ∗ pi ∗ r t o ∗ ( h c+h r ))+ dqdz / ( 2 ∗ pi ∗ k i n s ( i ) ) ∗ . ke ) . i f abs ( Tins new−T i n s o l d )<=1 Tins new1=Tins new . r t o . 9 Tanulus=Tan ( T i n s o l d . t . log ( r i n s / r t o ) . % hc from eq . TAWithIns .140 APPENDIX B. q1 . r i n s .%cp % 1 cp = 1488 l b / f t −s e c o n d = 1488∗3600 . h r . r i n s . e p s t o . r c i . i t e r =0. T i n s o l d=T i n s o l d +0. rh . r c i .7∗( abs ( Tins new−T i n s o l d ) ) . TAWithIns ) .%R visAn=v i s c o s i t y A n ( Tanulus ) .
h c=hc ( khc . t . Pr1=Pr ( kha . Part2=r t o ∗Uto∗ ftD ∗Tto . TAWithIns . Tto=T i n s o l d+dqdz ∗ log ( r i n s / r t o ) / ( 2 ∗ pi ∗ k i n s ( i ) ) . % TOTAL HEAT LOSS Ql=Qls ∗ PipeLengthWithIns . % f ( t ) from Ramey and W i l l h i t e ftD = FTIME( Uto . ke ) . r c i .%lbm/ f t ˆ3 Gr1=Gr ( T i n s o l d . visAn ) . end Qls=2∗pi ∗ r t o ∗Uto ∗ ( TbWithIns−TAWithIns ) . kha=kha1 ( Tanulus ) . h r . Part3=ke . 8 Uto=UtoCal ( h c . den=densityAn ( Tanulus ) . Th old =( Part1+Part2 ) / ( Part3+Part4 ) . % BTU InjLengthplot P lo tQ l =(1: PipeLengthWithIns ) ’ . Pr1 . kha ) . rh . Beta Gr .2. visAn . r c i . r i n s . =(Qls ) . q1 . r i n s ) . khc=khc1 ( Gr1 . ∗ I n j L e n g t h p l o t . Tins new=Tto−( r t o ∗Uto ) ∗ ( Tf−Th old ) ∗ log ( r i n s / r t o ) / ( k i n s ( i ) ) . g ) . HEAT LOSSES FROM SEA PART 141 visAn = visAn ∗ 1 4 8 8 ∗ 3 6 0 0 . den . r t o . r i n s . figure (1) .B. % from eq . Part4=r t o ∗Uto∗ ftD . k i n s ( i ) ) . Part1=TAWithIns∗ ke .
set ( gcf . 2 .% f t r r i s e r i =0. / 2 4 ) . legend ( l a b e l s . ’ n o r m a l i z e d ’ ) . ’ L o c a t i o n ’ . l a b e l s { i } = [ ’ \ lambda = ’ num2str ( LambdaValStr ) ] . . = r riseri . ’ Fontname ’ . 1 6 ) ylabel ( ’ Depth ( f t ) ’ ) LambdaValStr = s p r i n t f ( ’ %0. ’ F o n t S i z e ’ .1458. . . rto rci rco rh = 0. [ 0 0 1 1 ] ) . ’ XAxisLocation ’ . I n j L e n g t h p l o t . ’ YDir ’ . l a m d a i n s W i t h I n s ( i ) . ’NE ’ ) set ( gcf . : ) . May 9 th . Fidan . CODES FOR HEAT LOSS CALCULATIONS plot ( PlotQl . end end Code block: SeaPartwithoutIns. m o d i f i e d d a t e 2 0 1 1 . num2str ( PipeLengthWithIns ) ’ f t ’ ] . . c o l o r s ( i . % % % Heat l o s s c a l c u l a t i o n f o r w i t h o u t i n s u l a t i o n c a s e i s done clc . close a l l .142 APPENDIX B. ’ r e v ’ ) xlabel ( [ ’ Cumulative Heat L o s s e s ( i n BTU) f o r ’ . ’ c o l o r ’ . .3G ’ .m: % S o l u t i o n f o r O f f s h o r e p a r t with i n s u l a t i o n m a t e r i l a s % % We a r e l o o k i n g f o r t h e s o l u t i o n a t s t e a d y s t a t e c o n d i t i o n s % w r i t t e n 2 0 0 9 . ’ Li n e wi d th ’ . clear a l l . ’ C a l i b r i ’ . ’ O u t e r P o s i t i o n ’ . = 1. ’ top ’ . = r risero . % ft % ft % ft % ft . .60. ’ U n i t s ’ . S . r r i s e r o =18/12/2. 3 ) hold on grid on set ( gca .
%[ f t ˆ2/ hr ] epsci epsto epsriser = 0. = 70.2. S ) . 9 Tanulus=Tan ( Tto .077500155 [ f t ˆ2/ hr ] from q1=a l p h a 1 ∗ 3 8 7 5 0 .%F Beta Gr=BetaGr ( Tanulus ) .B. % c o n v e r s i o n f a c t o r t TbWithoutIns PipeLengthWithoutIns lamdaPipe Boltzman g =32.% i n [mˆ2/ s ] % 1 [mˆ2/ s ] =38750. 7 3 0 7 3 5 . TAWithoutIns ) . 15 h r=hr ( Tto . Tf = TbWithoutIns .9. Tgeo=TAWithoutIns .%cp % 1 cp = 1488 l b / f t −s e c o n d = 1488∗3600 l b / f t −hr = 21∗24. = ( 0 . r c i . % t h e r m a l c o n d u c t i v i t y o f water %i n [W/m K] ==1 Btu / ( h r f t ?F) = 1 . S=35. % ft % F T=(TAWithoutIns − 3 2 ) / 1 . % 50 m e t e r s % hrs % F % ft = 600/24.2∗3600∗3600. %[ Perry ’ s Chemical E n g i n e e r s ’ Handbook . = 0. 7 3 0 7 3 5 W/ ( mK ) k1 = SW Conductivity (T. HEAT LOSSES FROM SEA PART 143 rEaWithoutIns TAWithoutIns = 0. S ) .%F % h r from eq .%btu / hr−f t 2 −F % hc from eq .9. = 600. e p s c i . = 164. % c a l c u l a t e t h e o l d T i n s from eq . . Table 1 −4] ke=k1 / 1 . 0 7 7 5 0 0 1 5 5 .%R visAn=v i s c o s i t y A n ( Tanulus ) . e p s t o . r t o . = 0. 1 7 1 4 e −8). Boltzman ) . 16 Tto = TbWithoutIns .9. TAWithoutIns . 8 . a l p h a 1 = S W D i f f u s i v i t y (T. 7 th E d i t i o n .
’ O u t e r P o s i t i o n ’ . 3 ) % Create t i t l e t i t l e V a l S t r = s p r i n t f ( ’ %0.16) ylabel ( ’ Depth ( f t ) ’ ) set ( gcf . khc=khc1 ( Gr1 . ’ L i ne w id t h ’ . . ’ k ’ ) . TAWithoutIns . ’ top ’ . =(1: PipeLengthWithoutIns ) ’ . h r . visAn ) . ke ) . Ql ) . visAn . r c i . ’ C a l i b r i ’ . kha=kha1 ( Tanulus ) .3G ’ . hold on grid on set ( gca . ’ FontSize ’ . % TOTAL HEAT LOSS Ql=Qls ∗ PipeLengthWithoutIns . ∗ I n j L e n g t h p l o t . ’ FontSize ’ . ’ b ’ . . Beta Gr . ’ ft is = ’ titleValStr ’ i n Btu ’ ] . ’ Fontname ’ . q1 . ’ XAxisLocation ’ . rco . t . rh . I n j L e n g t h p l o t . t i t l e ( [ ’ Heat L o s s e s w i t h i n a l o n g t h e ’ num2str ( PipeLengthWithoutIns ) . kha ) . r t o ) . .%lbm/ f t ˆ3 Gr1=Gr ( Tto . r t o . . ’ r e v ’ ) xlabel ( ’ Cumulative Heat L o s s e s i n BTU ’ . Pr1=Pr ( kha . . r c i . . . den=densityAn ( Tanulus ) . ’ YDir ’ . =(Qls ) . . r t o . % from eq . % BTU InjLengthplot P lo tQ l figure (1) plot ( PlotQl . ke ) . CODES FOR HEAT LOSS CALCULATIONS visAn = visAn ∗ 1 4 8 8 ∗ 3 6 0 0 . g ) . 8 Uto=UtoCal ( h c . rh . % f ( t ) from Ramey and W i l l h i t e ftD = FTIME( Uto .144 APPENDIX B. [ 0 0 1 1 ] ) . ’ U n i t s ’ .14 . h c=hc ( khc . Pr1 . set ( gcf . ’ Color ’ . Qls=2∗pi ∗ r t o ∗Uto ∗ ( TbWithoutIns−TAWithoutIns ) . ’ n o r m a l i z e d ’ ) . den . .
5 lbm/ f t N−80 c a s i n g % The a n n u l u s c o n t a t i n s a s t a g n a n t g a s a t z e r o gauge p r e s s u r e a t w e l l h e a d % and c a s i n g i s cemented t o s u r f a c e i n a 12− i n h o l e . t u b i n g s e t on a p a c k e r i n 9 5/8 in . m o d i f i e d d a t e 2 0 1 1 . S . HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 145 B. % We a r e l o o k i n g f o r t h e s o l u t i o n a t s t e a d y s t a t e c o n d i t i o n s % w r i t t e n 2 0 0 9 . 5 i n . . close a l l . % % clc . . 2 C a l c u l a t i o n s o f Heat L o s s e s from S u r f a c e L i n e s % % 3 . 5 3 .3 Heat Losses from Sea Floor to Reservoir (a) Sea Floor to Reservoir with insulation.B. The following are the MatLab codes for our heat loss calculation starting from sea ﬂoor to reservoir with insulation and without insulation % Example 1 0 . clear a l l .3: Schematic representation of resistance to heat transfer sea ﬂoor to reservoir. Figure B. (b) Sea Floor to Reservoir without insulation.3. May 9 th % Fidan .
9. 9 6 ] . = 100. % ft % ft % ft % ft % ft % ft % Btu/ f t −hr−F % Btu/ f t −hr−F = [0. = 0.166 0. = 0.194 0. = 0.499 0. = 1000.5000. % STEP2 : C a l c u l a t e f ( tD ) a t t tD=aEWithIns ∗ InjTimeWithIns / ( rwWithIns ˆ 2 ) .388 0. = 0. CODES FOR HEAT LOSS CALCULATIONS roWithIns rinsWithIns rciWithIns rcoWithIns rwWithIns rEaWithIns aEWithIns einsWithIns eciWithIns lamda EWithIns lamda cemWithIns lamda insWithIns InjTimeWithIns TAWithIns TbWithIns PipeLengthWithIns = 0.1458. f=f t ( c . . ftD=f .2292. Rh2=Rh .146 APPENDIX B.4010.3556. = 24. = 21. c=2∗pi ∗Rh2∗ lamda EWithIns . = 0.96.569 0 . = 12. = 0. tD ) . = 0. CollectDatasInjIns =[]. B68 Up=0. = 600. % days % F % F % ft c o l o r s = l i n e s ( length ( l a m d a i n s W i t h I n s ) ) . f o r i =1: length ( l a m d a i n s W i t h I n s ) % STEP1 : I n i t i a l Assumption o f t h e T o t a l Thermal R e s i s t a n c e Rh=(log ( r i n s W i t h I n s / roWithIns ) / l a m d a i n s W i t h I n s ( i ) ) / pi .9. = 0. % STEP3 : C a l c u l a t e Tci from Eq .
.3. B. den =0. 4 5 . % STEP4 : C a l c u l a t e Tins from Eq . 9 2 lamda aa =(0. . .11∗10ˆ( −8)∗((1/ e i n s W i t h I n s )+( r i n s W i t h I n s / r c i W i t h I n s ) ∗ . ∗ ( Ui+Upi+Upw+Upo+Uins ) . 6 . ( 1 / e c i W i t h I n s −1))ˆ( −1)∗F+lamda aa / ( r i n s W i t h I n s ∗ . % lamda a = 0 . % Npr ˆ 0 . Ngr=(gc / g ) ∗ 7 . . 7 0 Ui =0.Upo=0. g =1. Uins=log ( r i n s W i t h I n s / roWithIns ) / l a m d a i n s W i t h I n s ( i ) .54∗10ˆ( −5)∗Tan + 0 . 3 1 2 . hcan =4.049∗ lamda a ∗Ngr ˆ 0 . . Tci=TAWithIns+(( TbWithIns−TAWithIns ) / ( 2 ∗ pi ∗Rh ) ) . Ban=1/(460+Tan ) . 3 3 3 ∗ 0 . 6 3 through B.Upw=0. Upi =0. 9 2 ) . lamda a = ( 2 . % STEP6 : C a l c u l a t e Rh u s i n g Eq .54∗10ˆ( −5)∗T+0. 0 1 6 4 . i f rEaWithIns >0 Uea=log ( rEaWithIns / rwWithIns ) / lamda EWithIns . 8 e −004)∗Tan + 0 . .076∗((460+60)/(460+Tan ) ) . 4 = 0 . 1 2 ∗ 1 0 ˆ 7 ∗ ( r c i W i t h I n s −r i n s W i t h I n s ) ˆ 3 ∗ g ∗ . % From FigB . . den ˆ2∗Ban ∗ ( Tins−Tci ) / ( gc ∗ v i s ˆ 2 ) . 4 1 v i s =2. end Uf=ftD / lamda EWithIns . .B. . gc =1. HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 147 Ucem=log ( rwWithIns / r co W i t h I n s ) / lamda cemWithIns . ∗ (Up+Ucem+Uea+Uf ) . %STEP5 : C a l c u l a t e hcan from Eq . Tins=TbWithIns −(( TbWithIns−TAWithIns ) / ( 2 ∗ pi ∗Rh ) ) . 6 6 Tan=( Tins+Tci ) / 2 . log ( r c i W i t h I n s / r i n s W i t h I n s ) ) . B. 1 0 . .0164 v i s =2. F=((460+ Tins )ˆ2+(460+ Tci )ˆ2)∗(920+ Tins+Tci ) . else Uea=0.
Rhc=1/(2∗ pi ) ∗ ( Ui+Upi+Upw+Upo+Uins+Ucan+Up+Ucem+Uea+Uf ) . ( Ui+Upi+Upw+Upo+Uins ) . 4 1 v i s =2. c=2∗pi ∗Rh2∗ lamda EWithIns . CODES FOR HEAT LOSS CALCULATIONS Ucan=1/( hcan ∗ r i n s W i t h I n s ) . Rh=Rhc . B. % STEP2 : C a l c u l a t e f ( tD ) a t t tD=aEWithIns ∗ InjTimeWithIns / ( rwWithIns ˆ 2 ) . . Uins=log ( r i n s W i t h I n s / roWithIns ) / l a m d a i n s W i t h I n s ( i ) .0164 . den =0.076∗((460+60)/(460+Tan ) ) .148 APPENDIX B. ∗ (Up+Ucem+Uea+Uf ) . . %STEP5 : C a l c u l a t e hcan from Eq . Upo=0. end Uf=ftD / lamda EWithIns . . while abs ( Rhc−Rh) > 1 e−4 c o u n t e r=c o u n t e r +1. Upi =0. 7 0 Ui =0. f=f t ( c . % STEP3 : C a l c u l a t e Tci from Eq . 6 3 through B. B68 Up=0. c o u n t e r =0. % STEP4 : C a l c u l a t e Tins from Eq . Ucem=log ( rwWithIns / r c o W i t h I n s ) / lamda cemWithIns . Tins=TbWithIns −(( TbWithIns−TAWithIns ) / ( 2 ∗ pi ∗Rh ) ) ∗ . . 6 6 Tan=( Tins+Tci ) / 2 . else Uea=0. B. Upw=0. tD ) . ftD=f . i f rEaWithIns >0 Uea=log ( rEaWithIns / rwWithIns ) / lamda EWithIns . Tci=TAWithIns+(( TbWithIns−TAWithIns ) / ( 2 ∗ pi ∗Rh ) ) .54∗10ˆ( −5)∗T+0. % From FigB . Rh2=Rh .
. hcan =4. Ngr=(gc / g ) ∗ 7 . 1 0 . [ Rh Qls Ql ] ] . I n j L e n g t h p l o t . log ( r c i W i t h I n s / r i n s W i t h I n s ) ) . : ) . g ∗ den ˆ2∗Ban ∗ ( Tins−Tci ) / ( gc ∗ v i s ˆ 2 ) . 3 3 3 ∗ 0 . ’ Li n e wi d th ’ . . num2str ( PipeLengthWithIns ) ’ f t ’ ] . ( rinsWithIns / rciWithIns ) ∗ . . . . end Rh=Rhc . ∗ I n j L e n g t h p l o t . lamda aa =(0.11∗10ˆ( −8)∗((1/ e i n s W i t h I n s ) + . ’ FontSize ’ . 0 1 6 4 . gc =1. % STEP6 : C a l c u l a t e Rh u s i n g Eq . % HEAT LOSS PER UNIT LENGTH Qls=(TbWithIns−TAWithIns ) /Rh . % TOTAL HEAT LOSS Ql=Qls ∗ PipeLengthWithIns . ’ Fontname ’ . ’ r e v ’ ) xlabel ( [ ’ Cumulative Heat L o s s e s ( i n BTU) f o r ’ . lamda a = ( 2 .B. . c o l o r s ( i . ( 1 / e c i W i t h I n s −1))ˆ( −1)∗F+lamda aa / ( r i n s W i t h I n s ∗ . . . 8 e −004)∗Tan + 0 . F=((460+ Tins )ˆ2+(460+ Tci )ˆ2)∗(920+ Tins+Tci ) . 16) =(1: PipeLengthWithIns ) ’ .049∗ lamda a ∗Ngr ˆ 0 . . Ban=1/(460+Tan ) . 9 2 ) . ’ XAxisLocation ’ . ’ YDir ’ . . . HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 149 v i s =2. InjLengthplot P lo tQ l figure (1) plot ( PlotQl . g =1. =(Qls ) . 3 1 2 . 6 Ucan=1/( hcan ∗ r i n s W i t h I n s ) . . ’ C a l i b r i ’ . C o l l e c t D a t a s I n j I n s =[ C o l l e c t D a t a s I n j I n s . 1 2 ∗ 1 0 ˆ 7 ∗ ( r c i W i t h I n s −r i n s W i t h I n s ) ˆ 3 ∗ . ’ top ’ .54∗10ˆ( −5)∗Tan + 0 . 3 ) hold on grid on set ( gca . ’ c o l o r ’ .3. Rhc=1/(2∗ pi ) ∗ ( Ui+Upi+Upw+Upo+Uins+Ucan+Up+Ucem+Uea+Uf ) . . .
= 0. clear a l l . ’ n o r m a l i z e d ’ ) . = 0.9. 5 3 . [ 0 0 1 1 ] ) . / 2 4 ) . = 12.96.m: % Example 1 0 . ’NE ’ ) set ( gcf . % % clc . 2 C a l c u l a t i o n s o f Heat L o s s e s from Wellbore % % 3 .1458. = 0. May 9 th % Fidan .3556. = 0. close a l l . = 0. l a b e l s { i } = [ ’ \ lambda = ’ num2str ( LambdaValStr ) ] .150 APPENDIX B. legend ( l a b e l s . t u b i n g s e t on a p a c k e r i n 9 5/8 in . roWithoutIns rciWithoutIns rcoWithoutIns rwWithoutIns rEaWithoutIns aEWithoutIns einsWithoutIns eciWithoutIns lamda EWithoutIns lamda cemWithoutIns InjTimeWithoutIns TAWithoutIns TbWithoutIns PipeLengthWithoutIns = 0. % Btu/ f t −hr−F % Btu/ f t −hr−F % days % F % F % ft % ft % ft % ft % ft % ft . CODES FOR HEAT LOSS CALCULATIONS ylabel ( ’ Depth ( f t ) ’ ) LambdaValStr = s p r i n t f ( ’ %0. = 0. 5 i n . = 600. ’ L o c a t i o n ’ .4010. ’ O u t e r P o s i t i o n ’ .9. S . m o d i f i e d d a t e 2 0 1 1 . set ( gcf . = 100. l a m d a i n s W i t h I n s ( i ) . 5 lbm/ f t N−80 c a s i n g % The a n n u l u s c o n t a t i n s a s t a g n a n t g a s a t z e r o gauge p r e s s u r e a t w e l l h e a d % and c a s i n g i s cemented t o s u r f a c e i n a 12− i n h o l e .3G ’ . = 0. = 24. end Code block: InjWellwithoutIns. = 1000. = 21. ’ U n i t s ’ . 2 .5000. % We a r e l o o k i n g f o r t h e s o l u t i o n a t s t e a d y s t a t e c o n d i t i o n s % w r i t t e n 2 0 0 9 . .
t . % from eq . % s t e p 4 : c a l c u l a t e t h e o l d Th Th old =( r t o ∗Uto∗ ftD ∗ Tf+kcement ∗Tgeo ) / ( r t o ∗Uto∗ ftD+kcement ) . visAn khc=khc1 ( Gr1 .3. Beta Gr=BetaGr ( Tanulus ) . Pr1=Pr ( ) . dqdz=dqdz ( 1 ) . % hc from eq . e p s i n s . e p s c i . % s t e p 5 : c a l c u l a t e t h e new Tci . h c=hc ( khc . Boltzman ) . 15 h r=hr ( T i n s o l d . 1 5 . Beta Gr . r i n s . kha=kha1 ( Tanulus ) . 9 and % Uto from eq .%kha . g ) . rco . r c i . r t o . k i n s . r c i . q1 .B. % Step 1 a s s i g n random v a l u e f o r dqdz dqdz=randperm ( 1 0 0 ) . hc from eq . Pr1 . visAn . kha ) . % f ( t ) from Ramey and W i l l h i t e ftD = FTIME( Uto . ke ) . HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 151 lamdaPipe =600. rh . 8 Uto=UtoCal ( h c . 9 Tanulus=Tan ( T i n s o l d . T c i o l d . r i n s ) . den=densityAn ( Tanulus ) . r c i . Gr1=Gr ( T i n s o l d . den . T c i o l d ) . r i n s . % c a l c u l a t e t h e o l d T i n s from eq . % Step 3 : h r from eq . kcement ) . visAn=v i s c o s i t y A n ( Tanulus ) . T c i o l d . 8 and f ( t ) from Ramey % h r from eq . 16 % Step 2 : Tci = Geothermal Temperature T c i o l d=TAWithoutIns . r i n s . h r . rh .
T c i o l d . dqdz=2∗pi ∗ r t o ∗Uto ∗ ( Tf−Th ) . r c i . i f abs ( Tci new−T c i o l d )<=1 Tci new1=Tci new . h c=hc ( khc . Part3=kcement . 8 . visAn . r i n s . khc=khc1 ( Gr1 . % from eq . % hc from eq . visAn=v i s c o s i t y A n ( Tanulus ) . Th=(Part1−Part2 ) / ( Part3−Part4 ) . e p s c i . hc from eq . % Step 3 : h r from eq . 8 and f ( t ) from Ramey % h r from eq . 9 and Uto %from eq . T c i o l d=T c i o l d +0. kha ) . Beta Gr . Part2=r t o ∗Uto∗ log ( rh / r c o ) ∗ Tto . kha=kha1 ( Tanulus ) . Gr1=Gr ( T i n s o l d . g ) .7∗( abs ( Tci new−T c i o l d ) ) . i t e r =0. Tci new=Tci new1 . Part4=r t o ∗Uto∗ log ( rh / r c o ) . 15 h r=hr ( T i n s o l d . den=densityAn ( Tanulus ) . Beta Gr=BetaGr ( Tanulus ) . r i n s . r c i . T i n s o l d=Tto −(( dqdz ∗ log ( r i n s / r t o ) ) / ( 2 ∗ pi ∗ k i n s ) ) . r i n s ) .152 APPENDIX B. Pr1 . 9 Tanulus=Tan ( T i n s o l d . T c i o l d . else while abs ( Tci new−T c i o l d )>=1e−4 i t e r=i t e r +1. T c i o l d ) . CODES FOR HEAT LOSS CALCULATIONS Tci new=Th old +(( r t o ∗Uto∗ log ( rh / r c o ) ) / ( kcement ) ) ∗ ( Tto−Th old ) . 1 5 . Boltzman ) . Part1=T c i o l d ∗ kcement . den . r c i . e p s i n s .
HEAT LOSSES FROM SEA FLOOR TO RESERVOIR 153 Uto=UtoCal ( h c . rh . rco . k i n s .B. kcement ) . % s t e p 5 : c a l c u l a t e t h e new Tci Tci new=Th old +(( r t o ∗Uto∗ log ( rh / r c o ) ) / ( kcement ) ) ∗ ( Tto−Th old ) . % s t e p 4 : c a l c u l a t e t h e o l d Th Th old =( r t o ∗Uto∗ ftD ∗ Tf+kcement ∗Tgeo ) / ( r t o ∗Uto∗ ftD+kcement ) . t . r i n s . ke ) . rh .3. h r . % f ( t ) from Ramey and W i l l h i t e ftD = FTIME( Uto . end end . r t o . q1 .
DimPipeIns = [ 0 . 6 7 1 . Diameter ) % Fidan . 4 7 0 . 7 8 8 . 5 6 6 . 1 112 134 158 186 217 252 290 334 4 1 . 3 7 2 . 4 8 3 .4 Table 14 from Prats [41] The following are the MatLab codes for our heat loss calculation for surface lines without insulation radiation and free convection number interpolated values using Table B. % I n t e r p o l a t i o n f o r the htc % P r a t s pg234 t a b l e B. 2 5 3 . 6 5 8 . 8 9 3 . 8 6 5 .14 from Prats[41]. 3 100 120 141 166 194 225 260 299 343 4 4 . 3 5 1 . 5 116 137 162 189 221 256 294 338 4 2 . 5 1 2 4 8 12 2 4 ] . 2 4 9 . 4 7 6 . Table14 . 4 6 9 . 4 7 5 . 9 5 9 . 9 104 124 146 171 198 230 265 304 348 4 6 . SurfTemp =[130 180 230 280 330 380 480 580 680 780 880 980 1080 1180 1 2 8 0 ] . 1 6 3 . 8 6 5 . 4 8 1 . 8 90 107 127 149 174 202 234 269 307 352 4 8 . return . 8 5 7 . 3 5 4 . S . 3 7 2 . 9 111 132 156 184 215 250 289 332 3 9 . 4 4 6 . 8 108 129 153 180 212 247 286 3 2 9 ] . 1 4 Table14 = [ 5 0 . CODES FOR HEAT LOSS CALCULATIONS B. 0 6 7 . 5 6 8 .154 APPENDIX B. 2 7 4 . 0 4 8 . Temp . 5 6 0 . 1 9 1 . 7 7 5 . DimPipeIns . 5 79 8 6 . 1 6 2 . 8 6 1 . ht=interp2 ( SurfTemp . 7 5 7 . 2 7 9 . 9 9 6 . function ht=htTable14 (Temp . Diameter ) . 2 5 1 .
802 .84 1.5∗ log ( tD ) + 0 .588 .96 1.54 2.00 1. 4 4 5 .02 1.644 .772 .35 1.572 .58 1. 3 1 3 . 3 1 8 .984 1.5.800 . 3 1 3 .439 . tD ) % Fidan .37 1.578 . 3 4 5 .39 2.87 1.17 2.44 2.77 1. 3 7 3 .25 1.04 1. function f=f t ( r .910 .616 .511 . 4 3 3 .58 2. R=[1000 500 100 50 20 10 5 2 1 .64 1. 3 2 3 .52 1.629 . 3 1 4 .59 1.17 2.568 .00 2. 0 1 0 ] .57 1.698 . end return . 4 1 7 .806 .B.38 2. r .48 1.22 1.11 1.790 . 4 3 8 .806 .20 1.97 1.424 . 3 1 1 .623 . 5 1 2 5 10 20 50 1 0 0 ] .03 1.59 1.86 1.37 1.89 2 . A.65 1.958 . tD ) .89 2.09 2.74 2. else f=interp2 (R. 2 . 9 0 ] .56 1.615 .86 1.430 .24 1.81 1.619 .81 2.15 2.427 .12 2.811 .01 1.473 .99 2.423 .89 2.75 2.02 1.02 1.37 2.872 . S .423 .17 2. 3 9 6 .452 .617 .38 1. 1 .05 1.77 2.08 1.87 1.95 1.86 2.614 .745 .88 1. 0 2 .16 2.842 .811 . . 1 .24 1.44 1.57 2.538 .57 2.801 .15 1. 0 5 .820 . 3 1 2 .421 .1 from Prats[41].73 2.42 2.67 1. i f tD>100 f =0. 4 0 3 .422 . t = [ .666 .88 2.89 2.73 2. F(TD) CALCULATION ALSO KNOWN AS RAMEY[?] 155 B.39 2.84 2. % I n t e r p o l a t i o n o f t h e v a l u e o f f ( t ) i n unsteady c o n d u c t i o n A= [ .97 1.34 1.48 2.803 .40 1.01 1.58 2.01 1.69 1.00 1.73 1. 3 3 0 .16 2.57 2.66 1.05 2.802 .24 1. 3 1 6 .71 2.59 1.94 1.72 2.40 2.51 2.66 1.5 f(tD) calculation also known as Ramey[42] The following are the MatLab codes for our heat loss calculation for sea ﬂoor to reservoir with/without insulation interpolated values using Table 10.63 1.56 2. 2 .36 1. t . 5 .
Appendix C Results for Diﬀerent Insulation Materials U singW hiteAerogelλW A = 0. 156 .0081BT U/(f t − hr −o F ) Figure C. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel.1: Steam temperature distribution for diﬀerent injection temperature vs depth (ft) .
Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. 1 year. Figure C. .157 Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel.3: Steam quality distribution for diﬀerent injection temperature vs depth (ft).2: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 1 year.
1 year.4: Heat loss distribution for diﬀerent injection temperature vs depth (ft).158 APPENDIX C.0162BT U/(f t − hr −o F ) Figure C. 1 year. RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C.5: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. U singF iberglassλF G = 0. .
Figure C. . 1 year. 1 year.7: Steam quality distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass.6: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass.159 Figure C.
U singCarbonF iberλCF = 0.8: Heat loss distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. .9: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). 1 year.160 APPENDIX C. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber.0208BT U/(f t − hr −o F ) Figure C. RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C.
.10: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber.11: Steam quality distribution for diﬀerent injection temperature vs depth (ft).161 Figure C. 1 year. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber.
1 year.162 APPENDIX C. U singT hermolasticInsulationλT I = 0. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber.0237BT U/(f t − hr −o F ) . RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C.12: Heat loss distribution for diﬀerent injection temperature vs depth (ft).
163 Figure C. .13: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation.14: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. Figure C. 1 year.
Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. Figure C.16: Heat loss distribution for diﬀerent injection temperature vs depth (ft).04BT U/(f t − hr −o F ) .15: Steam quality distribution for diﬀerent injection temperature vs depth (ft). RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C. 1 year. U singCalciumSilicateInsulationλCaSil = 0. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation.164 APPENDIX C.
Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate.18: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. .165 Figure C.17: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). 1 year. Figure C. 1 year.
1 year. Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate.20: Heat loss distribution for diﬀerent injection temperature vs depth (ft). Figure C. 1 year.19: Steam quality distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C.166 APPENDIX C. .
21: Steam temperature distribution for diﬀerent injection temperature vs depth (ft).167 U singW hiteAerogelλW A = 0. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. 1 year.0081BT U/(f t − hr −o F ) Figure C. .22: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. Figure C.
Figure C.24: Heat loss distribution for diﬀerent injection temperature vs depth (ft).23: Steam quality distribution for diﬀerent injection temperature vs depth (ft).0162BT U/(f t − hr −o F ) . RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C. U singF iberglassλF G = 0. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel. 1 year.168 APPENDIX C. Tm = 122 o F and injection rate 4850 lbm/hr with white aerogel.
169 Figure C. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. 1 year. Figure C.25: Steam temperature distribution for diﬀerent injection temperature vs depth (ft).26: Steam pressure distribution for diﬀerent injection temperature vs depth (ft).
Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. 1 year.28: Heat loss distribution for diﬀerent injection temperature vs depth (ft).27: Steam quality distribution for diﬀerent injection temperature vs depth (ft).170 APPENDIX C. RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with ﬁber glass. . 1 year.
171 U singCarbonF iberλCF = 0. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber.0208BT U/(f t − hr −o F ) Figure C.30: Steam pressure distribution for diﬀerent injection temperature vs depth (ft).29: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). 1 year. . Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. Figure C.
1 year.0237BT U/(f t − hr −o F ) . 1 year. Figure C. RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C.31: Steam quality distribution for diﬀerent injection temperature vs depth (ft). U singT hermolasticInsulationλT I = 0. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber. Tm = 122 o F and injection rate 4850 lbm/hr with carbon ﬁber.32: Heat loss distribution for diﬀerent injection temperature vs depth (ft).172 APPENDIX C.
.34: Steam pressure distribution for diﬀerent injection temperature vs depth (ft).173 Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation. 1 year.33: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). 1 year. Figure C.
04BT U/(f t − hr −o F ) . RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C. 1 year.36: Heat loss distribution for diﬀerent injection temperature vs depth (ft). Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation.174 APPENDIX C. U singCalciumSilicateInsulationλCaSil = 0.35: Steam quality distribution for diﬀerent injection temperature vs depth (ft). 1 year. Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with thermolastic insulation.
. 1 year. Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate.38: Steam pressure distribution for diﬀerent injection temperature vs depth (ft). 1 year.175 Figure C.37: Steam temperature distribution for diﬀerent injection temperature vs depth (ft). Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate.
1 year. .39: Steam quality distribution for diﬀerent injection temperature vs depth (ft).40: Heat loss distribution for diﬀerent injection temperature vs depth (ft). RESULTS FOR DIFFERENT INSULATION MATERIALS Figure C. Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate. Tm = 122 o F and injection rate 4850 lbm/hr with calcium silicate.176 APPENDIX C. 1 year. Figure C.
4(3). Hazim S.html. http://www. [2] Schematic view of the oﬀshore platform.com/software/feketeharmony/media/webhelp/HTMLFiles/ReferenceMaterial/CalculationsandCorrelations/PressureLossCalculations. and Burton D. www. 177 .php?id=phy141:lectures:35.sunysb.consrv. Ziels.htm. united states patent Vacuum 4512721. http://www. [6] Chen ChouMing Ayres.com/software/rta/media/webhelp/ctecorrelations. 1982.ca.fekete.Bibliography [1] Schematic view of one of the heat transfer mechanism: Convection. Alternative ﬂowpattern maps can improve pressuredrop calculations of the aziz et al.H.htm.fekete. http://www. multiphaseﬂow correlation.gov/dog/Pictureawell/PublishingImages /OFFSHOREPLATFORM2.patentgenius.ic.com/patent/4512721.jp. [3] Two phase ﬂow correlations : Modiﬁed Beggs and Brill two phase correlations. insulated steam injection Paul S. [4] Two phase ﬂow correlations with pressure loss calculation : iﬁed Beggs and Brill two phase correlations with Pressure ModLoss. 1989. www. SPE Production Engineering.edu/Class/phy141md/doku. [5] Abu AlSoof Nimat B AlNajjar. tubing.
D. Willman. [14] J. P.. [13] J. Ove Bratland.V. Multiphase Flow in Wells. part 1. [16] Jr. Cornelius B. 2010.W. Beggs. [12] Dr.W.F.B. 1987. J. The ﬂow assurance site multiphase ﬂow assurance. Gibson S.P. A study of twophase ﬂow in inclined pipes. Bradford. [9] Dominuque Monfrin Bernard M.T.E. 1961. description of . Journal of Petroleum Technology. Brill and H. Marlin failure analysis and redesign. Paper SPE 16723 presented at the SPE Annual Technical Conference and Exhibition held in Dallas. 1973. Emeraude vapeur: A steam pilot in an oﬀshore environment. John Wiley & Sons. Journal of Petroleum Technology. TX. D. Taylor D. Valleroy and L. P. pages 123–130. Powers. 2003. The Oil and Gas Journal.G. Pattillo J. Couderc.drbratland. D.D. http://www.H. Fritchie. Beggs and J. Bleakley.W. Brill and H. 13(7):681–690.D. Mukherjee. [10] Malcolm Alister Grant & Paul F Bixley. Quettier. Runberg A. Heat Transfer Handbook. 1999. 2730 September. TwoPhase Flow in Pipes. Geothermal Reservoir Engineering. Gosch P. [11] W. V.178 BIBLIOGRAPHY [7] H.W. 1964. second edition.com/PipeFlow2/chapter1. Adrian.J. Here are case histories of two thermal projects. [8] Kraus Allan D Bejan. Brill. 2011.html. Sharp C. Laboratory studies of oil recovery by steam injection. Society of Petroleum Engineers. [15] G.W. Academic Press. 25(5):607–617. Verpeaux and Lisette H.
A. 21(1):79–86. A comprehensive wellbore stream/water ﬂow model for steam injection and geothermal applications.com. [18] S. 5(4):184–193. [23] A. 1965. [25] P. Alkaya G. [19] J.BIBLIOGRAPHY 179 failure. [24] M. Shi. 1969. [20] Jerry P. Wellbore behaviour during saturated steam injection. Diaz B. P. Brown. [22] Alton R. Kabir. Society of Petroleum Engineers. Hasan and C. 2002. Department of Chemical Engineering. Journal of Canadian Petroleum Technology. Durlofsky K. Prediction of bottomhole conditions for wet steam injection wells. Denver.xeng. 58 October 2003.R. Holmgren. Colorado. Journal of Canadian Petroleum Technology. Journal of Petroleum Technology. ﬁrst edition. 1981. Oddie H. 2003.R. Fluid Flow and Heat Transfer in Wellbores. Aziz L. Farouq Ali. Driftﬂux modeling of multiphase ﬂow in wellbores.M. Texas. Experimental study of pressure gradients occurring during continuous twophase ﬂow in smalldiameter vertical conduits. 21(2):82–88. Earlougher Jr. Journal of Petroleum Technology. 2628 February 2002. [17] Robert C. Holst and D. Flock. 1982. University of Calgary. Holmes. http://www.H. Hagedorn and Kermit E. 21(5):527–534. 2006. [21] L. 1980. S. Fontanilla and Khalid Aziz. 1966. SPE Journal. X steam for matlab. Paper SPE 74528 presented at the IADC/SPE Drilling Conference. Master’s thesis. Some practical considerations in the design of steam injection wells. J. 17(4):475–484. SPE Annual Technical Conference and Exhibition. . Fontanilla.L. Dallas. A Mathematical Model for the Prediction of Wellbore Heat Loss and Pressure Drop in Steam Injection Wells.J.
[27] Jan Dirk Jansen. . 1992. Master’s thesis. Wellbore heat losses and casing temperatures during steam injection. Frankfurt am Main. [35] A. Society of Petroleum Engineers. J. New functional correlations for saturated steam properties. Modeling and Optimization of Oil and Gas Production Systems. John H. Twophase ﬂow. Govier. and N. second edition. [36] Syed M. R. 6th World Petroleum Congress. 1972.180 BIBLIOGRAPHY [26] Hans H. 1963. Zubair Mostafa H. Pressure drop in wells producing oil and gas. Stanford University. Basic Heat and Mass Transfer. 11(3). 2011. Mills. Huitt. Duns Jr. Advanced Production Engineering Lecture Notes. 1966. [33] D. June 19 . 18(9):1157–1162. Desalination and Water Treatment.W. Society of Petroleum Engineers. Vertical ﬂow of gas and liquid mixtures in wells. [31] K. A.F. Germany. Los Alamos Science. C. Ros.26. Journal of Canadian Petroleum Technology. Thermophysical properties of seawater: a review of existing correlations and data. 1966. George W. Sharqawy. [30] L. 1999. pages 26–35. Huygen and J. Zhiyue and Kamy Sepehrnoori. ThawFront Dynamics of SuperInsulated Wells in Cold Environments. pages 25–32. Department of Energy Resources Engineering. Liles. 2007. Prentice Hall. Leutwyler. L. Enhanced Oil Recovery. Liang. [28] H. 1963 . Casing temperature studies in steam injection wells. [29] and Maria Fogarasi Khalid Aziz. 2009. Journal of Petroleum Technology. [34] Cristovao Marques. American Petroleum Institute. Lienhard Va. Lake. 1989. [32] Miller Mark A.
1962. S. . Thermal Recovery. Satter. Gosch P. Dallas. Hemanta. 31(9):1198–1208. [44] P. Farouq Ali.D. 17(7):845–851. Brill J. Payne.H. D. Society of Petroleum Engineers. Beggs. Beggs. [46] O.A. Sarathi and D. Shoham. [41] M. Heat losses during ﬂow of steam down a wellbore. 1972. Society of Petroleum Engineers. 3(2):263–272. Wellbore heat transmission. Mechanistic Modeling of GasLiquid twophase Flow in Pipes. Texas. Journal of Petroleum Technology. [43] D.A. and H. 11(3):299– 315. 2005.W. Mukherjee. Empirical equations to predict ﬂow patterns in twophase inclined ﬂow. G. [45] A. Wellbore heat losses and pressure drop in steam injection. Journal of Petroleum Technology. Journal of Petroleum Technology. [39] Palmer C. Ramey JR. Evaluation of inclinedpipe. 1988.redesign. 24(2):139–144.C. SPE Production Engineering.G. Journal of Petroleum Technology. Ellis. Petroleum Reservoir Simulation. A study of multiphase ﬂow behavior in vertical wells. Gibson S. International Journal of Multiphase Flow.P. Pattillo R. 1986.F. Olsen.D. [40] Palmer C. Paper SPE 74529 presented at the IADC/SPE Drilling Conference. Applied Science Publishers. ﬁrst edition. [38] E.M. 1992. Fritchie Jr. twophase liquid holdup and pressureloss correlation using experimental data. G. K.J. and H. Brill J. Payne. 1965. 2628 February 2002. Marlin failure analysis and redesign: Part 2 .D. 1985. Prats.P.M.M. 1979. 14(4):436–440. Pacheco and S.BIBLIOGRAPHY 181 [37] Brill James P. [42] H.
Squier. Paper SPE 74530 presented at the IADC/SPE Drilling Conference. Overall heat transfer coeﬃcients in steam and hot water injection wells. Calculated temperature behavior of hotwater injection wells. .L. 2628 February 2002.P. D. Journal of Petroleum Technology. Dallas. 1962.C. [50] YuShu Wu and Karsten Pruess.W. and E.W. Willhite. 1990. Shah.P. Horne and P. 19(5):607–615. part 3. Texas. SPE Reservoir Engineering. Gosch. [48] P.J. 5(4):531–538. 1967. An analytical solution for wellbore heat transmission in layered formations (includes associated papers 23410 and 23411 ).182 BIBLIOGRAPHY [47] Smith D. Journal of Petroleum Technology.D.D. Sharp S. Marlin failure analysis and redesign. Pattillo J. vit completion with realtime monitoring. 14(4):427–435. D. [49] G. Dougherty.