Beruflich Dokumente
Kultur Dokumente
by
Ochiagha Victor Ananaba
2007
by
Ochiagha Victor Ananaba, B.Eng.
Thesis
Presented to the Faculty of the Graduate School of
The University of Texas at Austin
in Partial Fulfillment
of the Requirements
for the Degree of
Approved by
Supervising Committee:
Paul Bommer
Dedication
To God Almighty.
To my loving and supporting parents Sir Emeka & Lady Nnenne Ananaba.
To my siblings, Nnem, Ugochukwu, Ugwunwanyi, Ogbugo & Amah (Papa).
To the woman that will be my wife.
Acknowledgements
I specifically want to thank my supervisor Dr. Augusto Podio for his continuous
support and encouragement through this research project. Under his supervision I have
greatly improved my knowledge and skills in the areas of petroleum production engineering
and artificial lift systems. It is an honor to have him as my supervisor and to be his friend.
I wish to thank Dr. Paul Bommer for the time that he took inside his very busy
schedule to read and review my thesis.
I will not forget to thank our Lab. Technician and my friend Tony Bermudez whose
support in maintaining and constructing my laboratory models made certain that I finished
my experiments in good time with high levels of accuracy.
I wish to thank Glenn Banm, Harry Linnemeyer, Ehiwario M., Acholem K., Ojifini
R., Elekwachi K. and Don Sorrell who were there to help whenever I needed assistance.
My special thanks go to our amiable graduate coordinator Cheryl Kruzie. I would not
be in UT if not for her kind and honest counseling.
Finally I would like to thank the companies that supported this research, Echometer
Company, ConocoPhillips, Yates Petroleum and Chevron. The comments and suggestions
from James McCoy, Lynn Rowland, John Patterson and Gabriel Diaz helped in shaping my
research.
I worked with Renato Bohorquez in the early days of this research and it was great.
Ochiagha Victor Ananaba
December 2007
ABSTRACT
The re-design of the internal geometry of static down hole gas separators directly
affects the gas liquid separation performance.
This thesis describes experimental results obtained after changing the dip tube design
from the conventional straight design to a helical design. Typically, a static down hole gas
separator with a conventional straight dip tube design depends on gravity to induce density
difference in the flowing wellbore fluid which causes gas liquid separation to occur. Thus,
the device is known as a gravity driven down hole gas separator.
vi
This research compared the experimental results and visual observations from
gravity driven down hole gas separators to that of static down hole gas separators with
helical dip tube designs known as static centrifugal down hole gas separators.
The visual observations showed that not only did the driving mechanisms for gas
liquid separation inside static centrifugal down hole separators include gravity it also
incorporated other means such as induced centrifugal forces that greatly improved overall
gas liquid separation. The 6 inch/second threshold downward superficial liquid velocity
generally regarded as the industry rule of thumb for down hole gas separators was
increased to 10 inch/second. In field units this is a 200 BPD increase in liquid production.
This research also studied the effect of increasing outer diameter of gravity driven
down hole gas separators from 3inches (2.75 ID) to 4inches (3.75 ID). The results
showed that liquid handling capacity increased by over 90% due to favorable flow regimes
observed inside the separator. However, critical examination of gas liquid separation
performances of both 3 inch OD and 4 inch OD separators in terms of downward liquid
superficial velocity reveal that gas liquid separation results are similar. It was concluded
therefore that downward superficial liquid velocity is a reliable parameter in the design of
down hole gas separators and that all gravity driven separators regardless of separator
outer diameter will operate in similar fashion except at different liquid flow rates.
Bubble rise experiment performed in this research project gave a range of 1 100 cp
as region of applicability for the results discussed in this thesis.
vii
Table of Contents
Acknowledgment v
Abstract...vi
List of Tables ........................................................................................................ xii
List of Figures ...................................................................................................... xiii
CHAPTER 1
Introduction ..............................................................................................................1
1.1
OBJECTIVE ..........................................................................................1
1.2
1.2.1.2
Continuous Flow Down hole gas separator for
Progressive Cavity Pumps - Patent No 5902378 .........................17
1.2.2 ACTIVE TYPE CENTRIFUGAL DOWN HOLE GAS
SEPARATORS ...........................................................................20
1.2.2.1
1.2.2.2
1.2.2.4
Apparatus for separating gas and solids from well fluids Patent No 6382317 B1.................................................................24
CHAPTER 2
28
EXPERIMENTAL FACILITIES.........................................................28
2.3.1.2
viii
2.3.2.2
2.4
ECHOMETER-TWISTER ..........................................52
2.6.3.2
CHAPTER 3
55
ix
133
CHAPTER 5
145
CONCLUSIONS...............................................................................145
5.1.1 CONCLUSIONS FROM COMPARISONS OF GRAVITY
DRIVEN SEPARATORS AND STATIC CENTRIFUGAL GAS
SEPARATORS .........................................................................146
5.1.2 THE EFFECT OF INCREASING SEPARATOR OUTER
DIAMETER FOR GRAVITY DRIVEN SEPARATORS .......147
5.1.3 CONCLUSIONS FROM BUBBLE RISE EXPERIMENT ....149
5.2
5.3
Appendix A ..........................................................................................................152
Schematics of the Echometer Separators .............................................................152
Appendix B ..........................................................................................................155
Schematics of the Patterson Separators ...............................................................155
Appendix C ..........................................................................................................159
Original data files .................................................................................................159
Nomenclature .......................................................................................................160
Abbreviations .......................................................................................................161
References ............................................................................................................163
Vita .165
xi
List of Tables
Table 2-1 - Sample Excel Spreadsheet for continuous flow test.......................................... 41
Table 2-2 Echometer gas separators configuration ............................................................ 44
Table 2-3 Patterson Separator Configuration ...................................................................... 46
Table 4-1 Dimensions of bubble rise experiment apparatus ........................................... 134
Table 4-2 Fluid Properties used in bubble rise experiment ............................................. 135
Table 4-3 Test data for glycerin in association with water............................................... 138
Table 4-4 - Test data for corn syrup in association with water ....................................... 139
xii
List of Figures
Figure 1-1 - Centrifugal Separator(Kobylinski et al) ................................................................ 8
Figure 1-2 - Gas flow through centrifugal separator (Kobylinski et al) ................................ 9
Figure 1-3 - Reverse-flow separator (Kobylinski et al) .......................................................... 10
Figure 1-4 Collar-Size down hole gas separator (McCoy and Podio10)....................... 12
Figure 1-5- Down-hole gas separator (Patterson and Leonard11) ........................................ 14
Figure 1-6 - Jongbloed et al12 ..................................................................................................... 17
Figure 1-7 Static Centrifugal Separator by Obrejanu Marcel13 .......................................... 19
Figure 1-8 Invention by Bunnelle P14.................................................................................... 21
Figure 1-9 - Centrifugal Separator by Kobylnski et al ........................................................... 23
Figure 1-10 Invention by Powers Maston15 ......................................................................... 24
Figure 1-11 Invention by Delwin Cobb16 ............................................................................. 26
Figure 1-12 Cross section (3) in Figure 1-11 .................................................................... 26
Figure 2-1 Schematic of experimental test facility .............................................................. 30
Figure 2-2 Laboratory facility ................................................................................................. 30
Figure 2-3 Laboratory test well .............................................................................................. 31
Figure 2-4 Laboratory Well .................................................................................................... 32
Figure 2-5 Turbine flow meter and valve between pump and mixer ............................... 33
Figure 2-6 - - ITT Barton floco positive displacement meter .............................................. 34
Figure 2-7 - Fisher Porter Flow Rator tube............................................................................. 35
Figure 2-8 - Thermodynamic Omega Air Flow Meter .......................................................... 36
Figure 2-9 - Sample Performance plot for Patterson (3X1) in continuous flow ............... 42
Figure 2-10 Echometer (3 X1.5) gas separator design ....................................................... 45
Figure 2-11- Echometer entry port geometry ......................................................................... 45
Figure 2-12 Echometer (4X1.75) gas separator design ...................................................... 45
Figure 2-13 4 inch OD Patterson Separator Design .......................................................... 47
Figure 2-14 3 inch OD Patterson Separator Design .......................................................... 47
Figure 2-15 Twister Separator (Bohorquez) ........................................................................ 50
Figure 2-16 Twister Connection ............................................................................................ 50
xiii
Figure 2-17 Diagrammatic of the forces acting in a static centrifugal separator ............ 51
Figure 2-18 Echometer - Twister .......................................................................................... 52
Figure 2-19 Patterson - Twister ............................................................................................. 53
Figure 3-1- Twister results in field units .................................................................................. 56
Figure 3-2 - Twister result in terms of superficial velocities ................................................. 57
Figure 3-3 Echometer - Twister result in terms of superficial velocities ......................... 58
Figure 3-4 Echometer - Twister results in field units ......................................................... 59
Figure 3-5 - Patterson - Twister result in terms of superficial velocities ............................ 60
Figure 3-6 - Patterson - Twister results in field units............................................................. 61
Figure 3-7 Comparison of Echometer Twister and Echometer (3X1) results in terms
of superficial velocity .................................................................................................................. 64
Figure 3-8- Comparison of Echometer Twister and Echometer (3X1) results in Field
Units .............................................................................................................................................. 65
Figure 3-9 - Pressure Drop between the entry ports and pump intake for Echometer
Twister and Echometer (3X1); Casing Pressure (Pc) = 10 13psi ...................................... 66
Figure 3-10 Pressure measurements during the tests ......................................................... 67
Figure 3-11 Pressure drop for Echometer-Twister and Echometer (3X1) at constant
gas rates; Pc = 10 13 psi .......................................................................................................... 68
Figure 3-12- Pressure drop for Echometer-Twister and Echometer (3X1) at constant
liquid rates; Pc = 10 13 psi ...................................................................................................... 71
Figure 3-13 - Comparison of Patterson Twister and Patterson (3X1) results in terms of
superficial velocity ....................................................................................................................... 72
Figure 3-14 - Comparison of Echometer Twister and Echometer (3X1) results in Field
Units .............................................................................................................................................. 73
Figure 3-15 - Pressure Drop between the entry ports and pump intake for Patterson
Twister and Patterson (3X1) separators; Casing Pressure (Pc) = 10 13psi ...................... 74
Figure 3-16 Pressure drop for Patterson-Twister and Patterson (3X1) at constant gas
rates; Pc = 10 13 psi ................................................................................................................. 75
xiv
Figure 3-17 - Pressure drop for Patterson-Twister and Patterson (3X1) at constant liquid
rates; Pc = 10 13 psi ................................................................................................................. 76
Figure 3-18 Patterson-Twister (2 twits) ................................................................................. 77
Figure 3-19 - Patterson Twister (2 twists) results in terms of superficial velocities ....... 78
Figure 3-20 - Patterson Twister (2 twists) gas separator results in field units ................ 79
Figure 3-21 - Comparison of Patterson Twister (4 twists) and Patterson Twister (2
twists) results in superficial velocity terms ................................................................................ 80
Figure 3-22 - Comparison of Patterson Twister (4 twists) and Patterson Twister (2
twists) results in Field Units ........................................................................................................ 81
Figure 3-23 - Pressure drop between the entry ports and pump intake for Patterson
Twister 2 twists and 4 twists gas separators; Casing Pressure (Pc) = 10 -13 psi.................... 82
Figure 3-24 Comparison of results for all static centrifugal separators in terms of
superficial velocities .................................................................................................................... 84
Figure 3-25 - Comparison of results for all static centrifugal separators in field units ..... 85
Figure 3-26 - Pump Liquid Fraction for Static Centrifugal Separators at 10 in/sec ......... 86
Figure 3-27 - Echometer (4X1) and Echometer (3X1) results compared in field units ... 90
Figure 3-28 - Echometer (4X1.5) and Echometer (3X1.5) results compared in field units
........................................................................................................................................................ 91
Figure 3-29 - Echometer (4X1.75) results in field units ........................................................ 93
Figure 3-30 Comparison of results of all Echometer gas separators in terms of
superficial velocity ....................................................................................................................... 94
Figure 3-31 - Pump Liquid Fraction for Echometer Separators at 6 in/sec ...................... 96
Figure 3-32 - Pump Liquid Fraction for Echometer Separators at 10in/sec ..................... 97
Figure 3-33 Comparison of all Echometer 4 inch OD separator results in field units . 98
Figure 3-34 - Pressure drop between the entry ports and pump intake for Echometer 4
inch OD gas separators for 2 phase gas liquid flow; Casing Pressure (Pc) = 10 -13 psi. 100
Figure 3-35 - Pressure drop between the entry ports and pump intake all tested
Echometer gas separators; Casing Pressure (Pc) = 10 -13 psi ............................................ 101
Figure 3-36 - Patterson (4X1) and Patterson (3X1) results compared in field units ....... 103
xv
Figure 3-37- Patterson (4X1.5) and Patterson (3X1.5) results compared in field units .. 104
Figure 3-38 Patterson (4X1.75) results in field units ........................................................ 105
Figure 3-39 - Patterson (4X2) results in field units .............................................................. 106
Figure 3-40 - Comparison of results for all Patterson 3 inch OD and 4 inch OD
separators in superficial velocity terms .................................................................................. 107
Figure 3-41 - Pump Liquid Fraction for Patterson Separators at 6 in/sec ....................... 109
Figure 3-42 - Pump Liquid Fraction for Patterson Separators between 8 9 in/sec ..... 110
Figure 3-43 - Comparison of all Patterson 4 inch OD separator results in field units ... 112
Figure 3-44 Pressure Drop between the entry ports and pump intake for Patterson 4
inch OD gas separators; Casing Pressure (Pc) = 10 13psi ............................................... 113
Figure 3-45 - Pressure drop between the entry ports and pump intake for Patterson 4
inch OD gas separators at varying gas and liquid rates; Casing Pressure (Pc) = 10 -13psi
...................................................................................................................................................... 114
Figure 3-46 Comparison of results for Echometer (4X1.75) with 5dip tube and Echometer
(4X1.75) with 2 dip tube in superficial velocity terms ................................................................. 116
Figure 3-47 - Comparison of results for Echometer (4X1.75) with 5dip tube and Echometer
(4X1.75) with 2 dip tube in field units ...................................................................................... 117
Figure 3-48 Standing Valve Assembly ................................................................................ 119
Figure 3-49 SV joint Gas Separator Connection ........................................................... 119
Figure 3-50 Echometer (3X1) with SV result in terms of superficial velocities ........... 120
Figure 3-51 - Echometer (3X1) with SV result in field units .............................................. 121
Figure 3-52 Comparison of Echometer (3X1) with and without Standing Valve in terms
of superficial velocities ............................................................................................................. 122
Figure 3-53 - Comparison of Echometer (3X1) with and without Standing Valve in field
units ............................................................................................................................................. 123
Figure 3-54- Pressure drop between the entry ports and pump intake for Echometer
(3X1) with and without Standing Valve .................................................................................... 124
Figure 3-55 - Pressure drop between the entry ports and pump intake for Echometer
(3X1) with and without Standing Valve at varying gas and liquid rates; Pc = 10 -13 psi ... 125
xvi
Figure 3-56 Flow Regimes observed in the gas separator annular area
(courtesy Renato Bohorquez7)................................................................................................. 126
Figure 3-57 Flow regime map for the annular space of 3 inch OD gravity driven gas separators7 127
Figure 3-58 Flow regime map for the annular space of 4 inch OD gravity driven gas separators .. 128
Figure 3-59 Flow regime map for the Twister separator annlus7 ................................... 130
Figure 3-60 Flow regime for Patterson Twister and Echometer Twister static
centrifugal separators ................................................................................................................ 131
Figure 4-1 Schematic of Laboratory Constructed Apparatus for testing bubble rise velocity ......... 134
Figure 4-4 Glycerin Rheology test ....................................................................................... 136
Figure 4-5 Glycol Rheology test .......................................................................................... 136
Figure 4-6 Corn Syrup Rheology test.................................................................................. 137
Figure 4-7 Viscosity plot for Glycerin in association with water at room temperature
...................................................................................................................................................... 138
Figure 4-8 Viscosity plot for Corn Syrup in association with water at room temperature
...................................................................................................................................................... 139
Figure 4-9 Combined viscosity plots for glycerin and corn syrup in association with
water at room temperature ...................................................................................................... 140
Figure 4-10 Examples of bubble diameter sizes measured.142
Figure 4-11 Mean bubble rise velocities in stationary liquid in an annulus ................... 143
xvii
Chapter 1
Introduction
1.1
OBJECTIVE
Most wells producing from mature reservoirs use artificial lift methods for oil
and gas production. Common artificial lift methods include beam pumping,
progressive cavity pumping and electric submersible pumping. All the mentioned
artificial lift systems exhibit a common problem: Gas Interference
The presence of free gas in beam pumps (sucker rod pumps) prevents the
traveling valve from opening at the appropriate time interval during the downstroke.
This is caused by the high compressibility of gas in the pump barrel. The traveling
valve may eventually open when the gas inside the barrel has been compressed
enough to overcome the fluid load on the plunger. In such a case fluid pound occurs.
In extreme cases the peak pressure of the trapped gas on the downstroke is
insufficient to overcome the hydrostatic head of the traveling valve; then the pressure
is not reduced enough on the upstroke to allow the standing valve to open and admit
new fluid. Both valves are essential stuck at a closed position and the pump refuses to
pump. This extreme case is known as gas locking.
In progressive cavity pumps (PCP) the produced liquid lubricates the rotor and
the stator so as to reduce the heat caused by friction. The presence of free gas in the
produced fluid reduces the lubricating function of the produced fluid so that the rotor
and stator are in direct contact. Temperature increase due to the direct contact causes
damage to the pump. In other cases gas in the produced fluid in PCP may change the
chemical composition of the elastomer in the stator of the pump which further
complicates the problem.
Electric submersible pumps (ESP) are typically used to handle high liquid flow
rates. Significant volumes of gas entering the pump especially at low intake pressures
degrade the pump performance, and dramatically reduce the head produced by the ESP.
This may prevent the pumped liquid from reaching the surface. The ESP is composed of
a down hole motor which is connected to a seal section which in turn is connected
to a centrifugal pump. It is imperative that the motor be cooled by the produced fluid
passing the outer casing. In the event that large quantities of gas pass the motor, the heat
transfer from the motor to the produced fluid will be drastically reduced, potentially
causing motor damage by overheating.
In all cases - beam pumps, PCP and ESP the pump volumetric efficiency is
reduced by the presence of gas. To combat the problem of reduced volumetric efficiency
and system damage down hole gas separators are used in conjunction with down
hole pumps.
The sole purpose of down hole gas separators* is to prevent gas from entering
into down hole pumps, or to at least reduce the quantity of gas entering into the pump
to permissible ranges where the pump efficiency is still acceptable.
Unfortunately many gas separator designs have not yielded the desired efficiency.
The widely used poorboy gas separator which depends on gravity segregation to
separate gas from liquids has become synonymous with inefficiency.
*
Down hole gas separators will mean the same thing as gas separators throughout this thesis
A thorough literature review on the subject of gas separator design was done to
study previous designs and relevant applications. Sources of information included
published technical papers, patents and thesis reports by Lisguiski, Guzman and
Bohorquez.
The scope of the present work emphasized the effect of the internal geometry
and induction of centrifugal forces on gas separator performance.
1.2
LITERATURE REVIEW
velocity of 0.5ft/sec (6in/sec) inside the separator dip tube annular area. A downward
mixture velocity of 0.5ft/sec is generally accepted as being below the rising (slip) velocity
of gas in low viscosity fluids. Clegg2 and McCoy3 et al described the reasons for the
inefficiency of the commonest down-hole gas separator design the poorboy separator.
The reasons for the inefficiency of the Poorboy gas separator according to the author
included the high downward liquid velocity inside the Poorboy separator and size of its
dip tube ID which the author considered as too small in diameter. The small ID dip tube
often causes excessive pressure drop inside the separator. The Shell (Schmit Jongbloed)
gas anchor formula:
gas anchor efficiency =
100
(1 + C Pwf0.66 Vsl0.5 )
1
Pwf= intake pressure at the anchor; Vsl=downward superficial velocity of liquids; C = gas anchor
constant (usually 0.2 based on laboratory data)
described by Clegg1 showed that the performance of any given size and type of gas
separator is largely dependent on the intake pressure at the anchor and the downward
superficial velocity of the fluids in the anchor. An examination on the formula done by
the author reveled that at zero pressure and zero velocity the anchor/gas separator
efficiency is 100% and that at high velocities (greater than 0.5 ft/sec) inside the gas
separator the separation efficiency is poor. Pressures above 400psig also resulted in low
efficiencies. The author however cautioned that actual experiences indicate that
separation may be significantly greater than what the formula predicts. The uncertainty in
the equation emerged from the use of the constant C which represented other
important variables such as viscosity, gas bubble size and dispersion. Laboratory results
that were not published indicated a constant of 0.2. The author warned that the
U=
2 g ( 1 2 ) Rb2
9
2
Where U = terminal velocity, ft/sec; g = 32.17 ft/sec2; L = liquid density; Ib/ft3 g = gas density,
Ib/ft3; = liquid viscosity; Ibm/(ft-sec); Rb = bubble radius (ft)
The second step used the calculated terminal velocity to calculate the area of the gas
separator (also called Mud Anchor or MA) using equation 3.
0.00935 QL
U EV
AMA =
STB
; EV = Pump efficiency
D
These calculated values are inputted into the computer program explained in
their paper to generate relationships between (1) pressure effects on gas bubble velocity
over constant viscosity and temperature (2) gas bubble velocity and diameter of the mud
anchor over different liquid flow rates (3) dip tube diameter and pressure drop in the gas
anchor over different liquid flow rates (4) pressure drop and dip tube length as a
function of liquid rate.
The results showed that gas bubbles travelled faster in smaller OD mud anchors
larger dip tube diameters yielded the smallest pressure drop and longer dip tubes had the
largest pressure drops.
Experimental results from Lisugurski5, Guzman6 and Bohorquez7 however
dispute the orders of magnitude of the results from Campbell and Brimhall4. Field
results9 based on Lisugurskis
operate efficiently at rates which would require longer gas separator lengths if Campbell
and Brimhalls
concluded that gas bubbles especially during the up - stroke of a sucker rod pumping
system coalesce more readily and rise faster in smaller separator annular areas compared
to larger annular areas.
Kobylinski et al 8 described the design, development and laboratory testing of a
new rotary gas separator, Figure 1-1 and Figure 1-2. The rotary gas separator is an active-
type centrifugal separator. Laboratory and field comparison were conducted between the
Centrifugal separator and the passive-type Reverse-flow separator, Figure 1-3.
Laboratory tests were done using water and air as test fluids in continuous flow
condition. The Reverse-flow separator uses the gravity separation mechanism for gas
liquid separation. The Centrifugal separator achieved separation of gas and liquid by the
use of cyclone and vortex technology. The characteristics of this method identified by
the authors were that the separated liquid is concentrated in the vicinity of the wall of the
separator while the gas phase concentrates at the center of the system. The authors
stated that dimensioning of the separator should be based on the equation of the
trajectory of the gas bubbles; they added that a general equation that would cover the
turbulence arising in the process is not available. Kobylinski et al8 believe that since both
bubble dimensions and proportionality constant between gas and liquid velocities are
unknown from Stokess law for laminar flow(1), reliance on experimental work for
design optimization remains the only alternative. A detailed discussion on bubble
dynamics is analyzed in the paper.
The results from the field tests8 complemented the results from the laboratory
and led to a 95% average improvement in fluid production when results from the activetype centrifugal separators were compared to the passive-type reverse-flow separators in
tested wells. The dimensions of both the centrifugal and reverse-flow separators were
not given.
McCoy and Podio10 gave a detailed description of the Collar Size gas separator,
Figure 1-4. They emphasized a maximum pressure loss of PSI for friction loss in the
dip tube. The authors also highlighted the need to allow for sufficient space in the gas
10
separator annular area. According to the authors sufficient flow area should exist so that
the gas flow rate around the ports in the gas separator will allow liquid to flow or fall
into the gas separator annulus. The authors noted the necessity to balance the area
available for flow in the wellbore and that inside the gas separator. Decreasing the casing
annulus will result in increased upward gas velocity which when above 10 ft/second will
suspend some of the liquid and allow mist flow to occur. Another consequence of casing
annulus reduction and increase of gas velocity will be the prevention of liquid from
flowing into the gas separator annulus. The authors stressed the need for the use of large
ports. Large ports allow liquid from the casing to fall by gravity force into the gas
separator because the pressures inside and outside the large ports are the same.
Kobylinski et al8 in their paper also recommended that for a gas separator to operate
efficiently, it must ingest the two phase mixture with minimal pressure drop. This is
necessary to prevent additional gas breakout inside the separator. The Collar Size
separator10 had a total port area which was approximately four times the area inside the
gas separator. The gas separator length received special treatment by McCoy and Podio10
they suggested that the dip tube length extend at least 18 inches below the gas separator
inlet perforations (separator ports). They based their calculation on a gas rise velocity of
6in/sec (0.5ft/sec) and an average pumping speed of 10 strokes per minute which
translates to a pumping cycle time of 6 seconds. The authors also looked at eccentricity
of the separator. Earlier studies noted by the authors showed that liquid concentrates
where tubing is placed against the casing wall and thus advised that gas separators outer
diameter should contact the casing wall, see Figure 1-4.
In wells with some deviation McCoy and Podio10 advised that the separator
should be allowed to rest on the low side of the casing since gas tends to flow up on the
11
high side of the casing annulus by installing any tubing anchors at a distance of 60 to 90
feet shallower than the pump intake.
Figure 1-4 Collar-Size down hole gas separator (McCoy and Podio10)
Patterson and Leonard11 ran some field tests in coal-bed methane wells in
Wyoming with some changes in the down hole pump setting depth interval and for an
increase in gas separator OD. The authors noted that while the modifications were not
fully understood or tested with significant number of installations the improvements
observed warranted some discussion. The tests were conducted in two wells and are fully
described in the paper.
Patterson and Leonard11designed different gas separators used in the field tests in
a bid to achieve greater pump efficiency. The gas separators used in the tests had smaller
slot width and included vent holes and a baffle to facilitate the evolution of gas - Figure
12
1-5 . The concept according to the authors assumes that a smaller slot width will reduce
the amount of gas entering the gas separator and the vent holes will allow the gas that
enters to vent back to the casing. The slot sizes ranged from 0.3 wide by 6 long for the
3.5 OD gas separator (2 in number) to a 3/16 wide by 10 long for the 5.5 OD gas
separator (8 in number). The 5.5 OD gas separators also had three diameter holes
in the swedge (see Figure 1.1-5). Both separator designs had the same dip tube OD but
different dip tube lengths 2 inches difference. The 3.5 OD gas separator was 24 feet
long whereas the 5 OD gas separator was 26 feet long.
The test well , 43-26, had a 3.5 OD 8 long gas separator installed with a
Progressive Cavity Pump (PCP) at 1446 ft after a bucket test had been conducted.
After some months a 5.5OD gas separator was attached to the PCP in test well 43-26.
Although the well contained coal particles which got into the gas separator and starved
the pump intake some useful evaluations on the effect of increase in gas separator cross
sectional area were made from test well 43-26.
Due to gas separator design changes the inlet area of the 5.5 OD gas separator
design increased four times compared to the 3.5 OD gas separator design. The 5.5
OD gas separator annular area (gas separator annular area = gas separator ID dip tube
OD) increased by approximately 3 times over the 3.5 OD gas separator.
The field results11 showed that no gas was produced through the tubing when the
5.5 OD separator was run with the PCP in test well 43-26. Patterson and Leonard11
infer that the differences in inlet area and cross sectional area available for flow could
have had an impact on gas separation and would appear that some combination of these
differences has a grater influence on gas separation than only increasing the cross
sectional area. In another well test where a 4 OD gas separator with some modifications
to the entry slot area and separator length was compared to a 3.5 OD gas separator
13
efficiency in the same well. The authors observed that whereas the 3.5 OD gas
separator produced gas through the tubing the 4 OD gas separator did not. The authors
observed that the increase in annular area must have contributed to pump efficiency
improvement. They however speculated that the increase in length of the 4 OD
separator or the baffle design of the gas separator might also have aided to the
improvement. The authors suggested that more field tests be done and visual modeling
experiments be evaluated with different geometries and configurations to better
understand the reason(s) behind the improvements. Patterson and Leonard11 made other
related conclusions in the paper which dealt with; downward liquid velocity, essence of
vent holes, position of the inlet of the gas separator relative to the perforations and the
age old theory that placing the intake of the pump below the perforated interval creates
an effective natural gas anchor (gas separator).
14
Guzman6 experimentally determined that placing the gas separator inlet at about
3feet below the lowest perforation results in natural separation that yields total gas
liquid separation. A gas separator is not needed in such cases as long as the annular liquid
downward velocity is less than 6 inches per second.
Guzman6 also suggested that the ports area should be equal the gas separator
annular area so that the superficial liquid velocity does not control the flow regime inside
the separator. The use of vent holes in the design of Patterson and Leonard11 in the
experiments conducted by in continuous flow Guzman6 showed that the vent holes do
not improve gas separation. The author suggested the use of single row slots instead of
multiple rows. He however noted that for the decentralized wells the results might be
difficult to predict due to well eccentricity10.
Several centrifugal gas separators have been patented over the years. Most of the
patented arts require the invention have moving parts whereas some do not. The next
sections will initially describe the parts and mode of operation of patented static
centrifugal separators and finally do same for active patented centrifugal separators
1.2.1
15
conduit is centrally positioned in the housing, the conduit is provided with openings,
preferably near the side of the sheet metal helix facing the bottom of the housing (19),
the gas discharge conduit (17) and the sheet, communicates with the outside of the
housing through the opening (21) that is above the supply openings (20).
In a reported experimental arrangement in which the outer diameter of the
helical channels was 7.5 cm (gas anchor ID = 3 inches), mixtures of varying gas/oil
ratios were supplied to a gas anchor according to the invention. The quantity of oil
passing the separator was from 1 1.5 cu. meters per hour (151 BPD 226 BPD).
When fluid mixtures (dispersions) having gas oil ratios of between 5 and 20
were supplied, the gas/oil ratio of the mixture flowing through conduit (14) was less
than 0.01.
This invention has no moving parts.
16
1.2.1.2 Continuous Flow Down hole gas separator for Progressive Cavity Pumps
- Patent No 5902378
This apparatus invented by Obrejanu Marcel13 in 1999 is a gas separator which
can be attached to the suction of a down hole pump to remove gas from the liquid
being pumped prior to the liquid entering the pump inlet. The separator has an elongate
17
housing having an annular chamber with guides which direct the liquid gas mixture to
flow in an annular path from the inlet to the outlet end. During this flow centrifugal
forces act to displace the gas content to the central region from which it is removed via a
separate central gas outlet so that liquid delivered to the pump inlet is greatly reduced in
its gas content.
In operation, the separator is attached in a coaxial fashion via sub (14) to the
lower end of a progressive cavity pump. In a gaseous environment the liquid will contain
dissolved gases and will enter the chamber (15) under formation pressure through the
inlet ports (16). When the pump is operated the reduction in pressure as a result of the
pump suction will cause some of the dissolved gas to come out of solution. The gas
liquid mixture is drawn upwardly within the tubular housing (12) and upon encountering
the helical flights (20) is guided thereby to move in a helical path. The centrifugal forces
created in the liquid as a result of the helical flow act to reduce the gas content of the
peripherally outer region of the flow and increase the gas content of the central region of
the flow. The angular momentum created in the liquid flow by the flights (20) is
maintained as the liquid moves upwardly into the expansion chamber (23). In this
chamber the cross sectional area of the flow passage is expanded as a result of the
termination of the flights (20), the tapering and termination of the spindle (17), and the
outwards flare of the inner wall of the tubular housing (12), the combined effects of
these resulting in a marked reduction in pressure of the liquid flow thus enhancing the
gas separation effect. The centrifugal force in the rotating liquid is effective to confine
the separated gas to the axial region of the chamber which rises above the rounded top
end (22) of the spindle. The separated gas flow through the axial exit passage (26) to the
exterior of the sub (14) where they can be released into the well bore, or if so desired
delivered to the surface through a separate conduit.
18
This separator has no moving parts within the separating chamber. To force the
liquid into the chamber the separator depends on both hydrostatic head and the pressure
drawdown created by the action of the PCP mounted above. Another interesting feature
about the invention is that multiple separation chambers could be attached just below
(18) for a two stage separation process before the liquid enters into the pump.
This separator design is currently been manufactured in commercial quantity in
Canada.
19
1.2.2
20
upward movement of the discharged gas within a wellbore where the separated unit is
situated.
21
(100) which assists in filtering debris from the fluid mixture. From the intake ports, the
fluid mixture enters the inducer (48) which pressurizes the fluid mixture and supplies it
to the centrifugal separator (50) via transition region (52). The transition region, which is
designed to provide a uniform rate of change through in flow direction and velocity to
the fluid mixture, conveys the fluid mixture smoothly to the centrifugal separator while
minimizing pressure loss. At the outlet end of the transition region, the tangential
velocity of the fluid approaches angular velocity of the centrifugal separator vanes and
the axial velocity of the fluid approaches the flow through velocity of the apparatus.
Liquid gas separation occurs at the inlet of the centrifugal separator region and
continues throughout its length. The liquid section is supplied to the pump through (86),
while the separated gas is vented via gas vents (90) into the space between the well casing
and the separator.
The rotary motion to the inducer and extending vanes are supplied by an
attached motor (20)
22
23
24
The gas and solids are removed from the well fluids in two separate steps by two
separate spirals, one spiral for the gas (66) and a separate spiral for the solids (70). An
upper gas spiral is positioned below the openings (60) in the outer tubular housing (44)
and a separate lower spiral spaced axially from the upper gas spiral is provided for the
solids. The spirals are positioned in the annulus between the outer tubular housing and
the inner flow tube (46). The spirals provide a helical flow and are spaced axially from
each other at a distance. The gas accumulates in the swirl chamber (80) between the
spirals and is librated from the liquid. The gas normally exists as large bubbles through
an inner gas annulus (72). The liquid flows downwardly in a helical path to the solids
spiral.
The solids, such as sand are separated from the well fluid by the solids spiral and
fall by gravity into the mud anchor or other suitable collection area. The liquid then
flows upwardly in the flow tube (46) to be pumped for flow to a surface location.
This invention makes use of induced centrifugal motion and gravity to separate
gas and solids from wellbore fluids.
The invention has no moving parts.
25
26
The research reported in this thesis extended the experiments of Kobylinski et al8
and the theory behind the inventions of Jongbloed et al12 and Obrejanu13 in the sense of
using gravity, agitation and centrifugal forces as physical mechanisms to obtain improved
gas liquid separation. The difference between this research and Kobylinski et al8 is in
the use of different experimental procedure, experimental facilities and most importantly
that the centrifugal down hole gas separator must have no moving parts; it must be
static similar to the inventions of Jongbloed et al12 and Obrejanu13.
This research also investigated experimentally the points raised by Patterson and
Leonard11 in terms of the effect of the increase in the gas separator annular area and
improvement in pump efficiency. Visual observation as Patterson and Leonard11
suggested was used to capture the separation mechanism(s) in the production
engineering laboratory at University of Texas at Austin as described in the next chapter.
27
Chapter 2
Experimental Facility and Procedure
This chapter fully describes the facilities, equipment and procedure used in
acquiring laboratory data used throughout this research. The down hole gas separators
used for the purposes of this experimental study are described in detail.
2.1
EXPERIMENTAL FACILITIES
The separator designs were installed in a laboratory well model and tested over a
range of 120 900 BPD of water and air rates between 13 115 MSCFD. The input
into the experimental test system was water and air at pre - determined rates Qg and Qw;
the output from the system included the pressures at the entry ports, tubing pressure and
the gas flow-rate through the dip tube of the separator. The inputs and outputs are
combined in a mathematical model to calculate the pump liquid fraction (pump
efficiency) of the separator relative to particular input values.
This chapter describes the facilities at The University of Texas Production
Laboratory and the procedure used to input and acquire data.
2.2
production laboratory facility used for testing down-hole gas separator designs. The test
facility is a closed loop system with manually controlled valves for fluid flow control.
Water was pumped in a loop into and out of a 3 - phase separator into the well (Figure
28
2-3) Air was supplied to the system by a compressed air line. Water and air meet at the
mixer before entering into the well. The hoses lead the mixture from the manifold
through the casing perforations into the well. Water is returned to the 3 phase
separator through a return a line. Air that passes through the dip tube is carried with the
water into the 3 phase separator and the rest rises up the casing.
29
30
31
2.3
of the gas and liquid phases inside the separator. Figure 2-3 shows close up pictures of
the laboratory well. Figure 2-4 shows the full laboratory well picture, notice that the
down hole gas separator is placed below the down hole pump. All the laboratory
tests were conducted with the gas separator situated in such position. The down-hole gas
separator components are positioned in the laboratory well as they would in a real well.
The mud anchor is the outer barrel of the separator. The mud anchor entry ports or
inlets allow water and some of the air to flow into the separator. The dip tube is the
small diameter tube inside the separator. The water flows down in the separator annular
area to the dip tube suction. Then the water flows up through the dip tube to the tubing
intake shown in Figure 2-4.
The bottom part of the casing has an ID of 6 inches. The upper part is PVC pipe
that extends to the rooftop of the Petroleum Engineering building at the University of
Texas, approximately 80ft. as seen in Figure 2-4. The bottom section of the casing has
several perforations, 31/64 inch in diameter, distributed at different positions. This way
it is possible to vary the relative location of the down-hole separator entry ports with
respect to the perforations.
32
Tubing Pressure
gauge (P3)
Ports Pressure
gauge (P2)
Location of the
separator
33
2.3.1
LABORATORY INSTRUMENTS
The instruments used in the conducting all the tests used for this research are
shown below. The functions that they performed are also explained.
2.3.1.1 LIQUID FLOW MEASUREMENTS
Figure 2-5 is a photo of the Daniel MRT97 turbine flow meter, used to measure
the water flow rate, installed in the liquid loop before the mixer. The water flow was
controlled by the valve in the same picture.
Figure 2-5 Turbine flow meter and valve between pump and mixer
The ITT Barton Floco positive displacement meter (ITT Barton, model 308K)
was used only for reference. It is installed between the turbine flow meter and the mixer.
34
35
36
37
in psig for positive values of pressure and in inches of mercury for vacuum. One of the
applications this pressure is to determine the pressure drop that occurs between the
separator in-take (P2) and the tubing pressure/discharge pressure (P3).
2.4
EXPERIMENTAL PROCEDURE
Use Figure 2-1 to follow the step - by - step procedure shown next.
Before beginning
1. Make sure that there is sufficient water in the separator using the level
control. Add water if necessary
2. Make sure that the desired ports in the manifold are open to inject flow from
the desired position relative to the down-hole gas separator entry ports
Starting the flow of fluids in the loop and setting the system in steady state
3. Turn on the pump
4. Use valve G to regulate the water flow rate. The gallons per minute read
by the turbine flow meter should approximately result in the desired BPD.
Valve G and the turbine flow meter are shown in Figure 2-5.
5. Gradually open the air flow to the desired percentage. It is usually set at
0% for the first experiment. There are three valves involved in the airflow.
First open valve D to let air in from the compressed air line. Then set the
desired percentage with valve E (seen in Figure 2-7).
38
6. Finally open valve F to let air mix with the water. Valve F should be
opened carefully. Otherwise the sudden injection of air can cause the
water to come out the top of the well.
7. By closing valve A, let water accumulate in the well until a desired bottom
hole pressure is obtained. The pressure gauge labeled BHP in Figure 2-1
indicates the bottom hole pressure. All the continuous flow tests were
run between 10 13 psi. Once the desired hydrostatic head is obtained,
regulate the flow out of the well to match the flow entering the well using
valve A. This way, the BHP and liquid level inside the three phase
separator are kept constant. This control is done throughout the test.
2.5
The plots are presented both in terms of oil field units and in terms of superficial
velocities.
In terms of oil filed units the x axis represents the input liquid flow rate in
BPD entering into the well through the perforations; the y axis is the gas flow rate in
MSCFD entering the laboratory well; the z axis is the gas rate through the separator in
MSCFD. This represents the gas that would enter the pump in a real well having the
down hole gas separator installed immediately below the pump intake.
In terms of superficial velocity the x axis is labeled the superficial liquid
velocity inside the separator in inch/second. The y axis represents the superficial gas
39
velocity inside the casing annulus in inch/second and the z axis is the gas rate through
the separator in MSCFD.
The height of each dot (and/or the vertical bar) on the 3 D performance plot
corresponds to the gas rate through the separator for a given liquid and gas rate either in
terms of oil filed units or in terms of superficial velocity.
The data used in plotting the performance plots are managed with an Excel
spreadsheet, see a sample data set in Table 2-1.
40
41
42
Figure 2-9 - Sample Performance plot for Patterson (3X1) in continuous flow
43
The naming procedure is given as: (separator name) (separator OD x dip tube OD).
Continuous flow tests2 were run on these separators between spring 2006 and
summer 2007 for the purposes of:
the continuous flow tests were conducted with fluid entering from below the separator through the
bottom four perforations located adjacent the separator see Figure 2-4
44
is 16 inches below the thread. There is a distance of 24 inches between the lower slots and
the dip tube suction and 44 inches between the upper slots and the dip tube suction.
Both the 3inch and 4inch OD Echometer separators have a wall thickness of 0.125
inch so that the IDs are 2.75 inch and 3.75 inch respectively. The dip tube wall thickness is
also 0.125 inch, making the ID of the dip tube a quarter of an inch less than the dip tube
OD.
The 4 inch OD Echometer separators have the same design configurations as the 3
inch OD Echometer separators, except for larger diameter.
A summary of the Echometer separator design configurations is shown in Table 2-2
below.
Table 2-2 Echometer gas separators configuration
Separator
type
Echometer
(3X1)
Echometer
(3X1.5)
Echometer
(4X1)
Echometer
(4X1.5)
Echometer
(4X1.75)
Number of
slots
Size of slots
(WXL)
(inch)
Total area of
slots (in2)
Area of
separator
dip tube
annulus (in2)
Area of
casing
separator
annulus (in2)
2X4
32
5.15
21.20
2X4
32
4.17
21.20
2X4
32
10.26
15.70
2X4
32
9.28
15.70
2X4
32
8.64
15.70
45
46
2.6.2
The naming principle is same as the Echometer designs. Continuous flow tests were
run concurrently for both designs between spring 2006 and fall 2007. The purpose of the
tests is same as listed for Echometer design.
The Patterson design has 16 thin and long entry slots. The slots are 1/8 inch wide
and 8 inch long. There are 0.5 in diameter vent holes. Table 2-3 is a summary of the
Patterson separator configuration.
Table 2-3 Patterson Separator Configuration
Separator
type
Patterson
(3X1)
Patterson
(3X1.5)
Patterson
(4X1)
Patterson
(4X1.5)
Patterson
(4X1.75)
Patterson
(4X2)
Number
of slots
Number
of inch
holes
Size of
slots
(WXL)
(inch)
Total area
of slots
(in2)
Area of
separator
dip tube
annulus
(in2)
Area of
casing
separator
annulus
(in2)
16
1/8 X 8
16
5.15
21.20
16
1/8 X 8
16
4.17
21.20
16
1/8 X 8
16
10.26
15.70
16
1/8 X 8
16
9.28
15.70
16
1/8 X 8
16
8.64
15.70
16
1/8 X 8
16
7.90
15.70
47
2.6.3
TWISTER
The separator named The Twister is the first in the series of static centrifugal
separators constructed since summer 2006. The initial results of the performance of the
Twister as reported by Bohorquez7 pointed to the need for more inquests into the
48
performance of static centrifugal separators based on the Echometer and Patterson separator
entry port designs.
The twister design uses a wire reinforced PVC hose used as a dip tube. The hose has
a 1.028 in OD and a 0.75 in. ID. The reinforced PVC is spiraled four full turns inside the
gas separator, Figure 2-15. The hose is twirled inside the gas separator and a plate is used to
secure the hose in place.
The straight dip tubes (for example, Echometer (3X1)) used for gravity separator
designs are directly connected to the wells tubing. But for the centrifugal design the
connection to the tubing is different. Figure 2-15 is a picture of the twister connection. The
arrows show the flow path for the gas through the gas vents and the liquid through the spiral
tube connection.
In operation the gas liquid mixture enters through the three circular entry ports of
the Twister separator. The helical dip tube induces a centrifugal motion on the mixture
entering through the ports. Gas is evolved and a coalescing zone is formed. The length of
the coalescing zone depends on the gas and liquid flow rates. The liquid mass is forced to
the inner walls of the separator by centrifugal forces and the gas mass accumulates at the
center (coalescing zone). While the gas rises to the gas vents at the separator connection to
escape into the casing annulus the liquid flows down towards the dip tube suction and
thereafter into the pump by gravity forces. Figure 2-17 shows a diagrammatic of the forces
acting on the mixture as soon as it enters into the separator annulus.
Laboratory observations show that the helical dip tube induces the centrifugal
motion by virtue of its design; a bubble coalescing zone is formed in the center of the
separator; the bubbles coalesce and become bigger bubbles in the coalescing zone and thus
rise faster; the liquid momentum is reduced by the helical nature of the dip tube. This greatly
The full construction detail for the Twister is covered in the thesis report by Bohorquez, 2006.
49
improved the gas pathway through the core of the separator eliminating the need for a gas
venting tube.
The operation of the twister is similar to the invention by Jongbloed et al12 since
both are static type separators. The basic difference between the two is that the twister does
not have a gas discharge conduit instead it has inclined gas vent holes at the connection head
(Figure 2-16). The Twister design also has similarity with the invention by Obrejanu13. Apart
from both separator designs been static by construction both separator designs depend on
the helical nature of the separator internal design to induce centrifugal motion and thus
centrifugal forces on the fluid flow inside the separator. The centrifugal forces induced
become the driving mechanism for gas liquid separation. The main differences between
the two are the entry port placement and gravity effects. The Twister has three entry port
circles at the top of the separator; the separator design by Obrejanu13 has the entry port at
the bottom part of the separator. The Twister depends to a significant extent on gravity for
gas liquid separation as well as on hydrostatic head to flow the mixture through the
separator. The gas separator by Obrejanu13 depends on both hydrostatic head and pressure
drawdown created by the action of the PCP to operate efficiently. Whilst the advantages of
having the entry ports at the bottom part of the gas separator is founded by density
difference; in the case of fines production the PCP rotor will erode at a faster pace causing
pre mature pump damage.
50
Gas Vent
holes
Spiral Tube
Connection
51
coalescing zone
Induced
centrifugal
motion
52
PATTERSON TWISTER
The Patterson Twister is the third in the series of static centrifugal separators
constructed. This separator design (Figure 2-19) has the same entry port geometry has the
Patterson (3X1). Like the Echometer Twister it was constructed to comparatively study
the effect of centrifugal forces on the previously constructed Patterson (3X1) separator
design. The Patterson Twister also has a dip tube with four full turns/twists with pitch
length of 12 14 inches lying at an angle of 45o on the inner walls of the separator.
53
54
Chapter 3
Analysis Of Experimental Results
3.1
55
The performance plots for each of the static centrifugal separators are presented in
the following sections and a comparative analysis is presented thereafter so that the effects
of the change in dip tube design are effectively captured.
3.1.1
continuous flow in both field units and in terms of superficial velocity. Notice that the
Twister achieved a zero gas flow rate through the separator up to a downward superficial
liquid velocity (Vsl) of 10 in/sec for all gas rates tested. This liquid rate is equivalent to 430
BPD. The area highlighted in red represents the optimum performance for the Twister
separator. Approximately no gas entered into the dip tube suction in this area for these
conditions of liquid and gas flow.
56
57
In Figure 3-3 and Figure 3-4 the performance results for EchometerTwister
separator are shown.
58
59
60
3.1.3
Both Figure 3-5 and Figure 3-6 show the results for the PattersonTwister Separator
in continuous flow in terms of superficial velocities and in field units.
61
62
3.2
static centrifugal separators in laboratory continuous flow experiments. For critical analysis,
comparisons are made between gas separators that have the same entry port geometry and
exact dip tube outer and inner diameters. Section 3.2 is organized as follows:
Section 3.2.3 studies how the number of twists inside a static centrifugal
separator affects performance. The performance of Patterson-Twister gas
separator with 4 twists is compared to Patterson-Twister gas separator with 2
twists.
3.2.1
Section 3.2.4 compares all the results for static centrifugal separators.
compared in terms of superficial velocities in Figure 3-7 in and in filed units in Figure 3-8.
The test points in the performance plots are red for the Echometer-Twister and black for
Echometer (3X1). As shown in the superficial velocity performance plot, Echometer (3X1)
63
separates all the air entering the test well for downward superficial liquid velocities below 7
in/sec, and the Echometer-Twister for downward superficial liquid velocities below 9.5
in/sec.
The plots in Figure 3-7 shows that the 6 inch/sec rule of thumb threshold for
downward superficial liquid velocity (Vsl) inside the separator for gravity driven separators
(green highlighted region) is surpassed. The limiting downward Vsl for optimum
performance inside the Echometer Twister gas separator design is 9.5 inch/sec; a 58%
increase over the conventional rule of thumb. In Figure 3-8, 9.5inch/sec optimum
downward Vsl for EchometerTwister gas separator design is equivalent to 480 BPD of
water. This represents a 200BPD increase over the optimum operational area for the
Echometer (3X1) gas separator design. The 7in/sec optimum downward liquid Vsl for the
Echometer (3X1) design marginally exceeds the rule of the thumb. This is an additional 40
BPD gas free liquid production.
64
Echometer Twister
Echometer (3X1)
Figure 3-7 Comparison of Echometer Twister and Echometer (3X1) results in terms of
superficial velocity
65
Echometer Twister
Echometer (3X1)
Figure 3-8- Comparison of Echometer Twister and Echometer (3X1) results in Field Units
The design of gas separators is a trade off between optimizing the separator annular
area and the pressure drop in the separator. Pressure drop analysis is performed on both gas
separator designs to determine the pressure difference between the entry ports, P1 and the
66
Pressure Drop for one phase liquid flow in Echometer (3X1) and Echometer-Twister
7
0
0
100
200
300
400
500
600
700
800
Echometer 3X1
Figure 3-9 - Pressure Drop between the entry ports and pump intake for Echometer
Twister and Echometer (3X1); Casing Pressure (Pc) = 10 13psi
67
68
Figure 3-11 and Figure 3-12 studies the effect of 2 phase gas and liquid flow on the
pressure drop values for the Echometer (3X1) and Echometer-Twister.
Figure 3-11 presents the pressure drop values for varying liquid rates and several
constant gas rates. The pressure drop values along constant gas rates are black for
Echometer (3X1) and red for Echometer-Twister. The plot shows that the pressure drop
across the separator is liquid rate dependent for both gas separator designs. The pressure
drop increases as the liquid rate increases in both gas separators.
Pressure Drop Analysis for Echometer (3X1) and Echometer-Twister at Constant Gas Rates
12
10
E 14MSCFD
E 41MSCFD
E 79MSCFD
E 115MSCFD
ET 14MSCFD
ET 39MSCFD
ET 76MSCFD
ET 110MSCFD
E = Echometer (3X1)
ET = Echometer Twister
0
0
100
200
300
400
500
600
700
Figure 3-11 Pressure drop for Echometer-Twister and Echometer (3X1) at constant gas
rates; Pc = 10 13 psi
It is usually expected that pressure drop values across the separator increase with
increasing dip tube length4,5. At 61, the dip tube length of Echometer-Twister is ft longer
69
than that of Echometer (3X1). Pressure drop analysis plot in Figure 3-10 agrees with the
literature but not in total. It is seen that between 450 BPD and 500 BPD the pressure drop
values for the Echometer-Twister are lower than that of Echometer (3X1). A quick
reference to Figure 3-8 will reveal that within the 450 500 BPD region, EchometerTwister separates all the air entering into the test well, so that only liquid enters into the
tubing return line3 .
The presence of 1-phase liquid flow inside the separator reduced the pressure drop values.
To explain this phenomenon a sketch of a section of a gravity driven separator under two
phase flow conditions is shown below
70
forces posed by liquid resistance to upward gas flow. The drag forces are shown as tiny
arrows pointing in the direction opposite to gas flow.
Pressure increases as the work done to overcome drag forces increases. This causes
pressure drop across the gas separator to increase during two phase gas liquid flow.
The additional pressure drop caused by phase interaction is absent for single phase
liquid flow conditions inside the separator. This is the basis for lower pressure drop values
for single phase liquid flow across the separator.
This explains why Echometer-Twister had pressure drop values lower than
Echometer (3X1) at rates of 450 500 BPD where single phase liquid flow was the
controlling flow regime inside the Echometer-Twister gas separator. At those rates the
controlling flow regime inside Echometer (3X1) was two-phase gas-liquid flow. Figure 3-8
distinguishes the single phase flow region for Echometer (3X1) as the green highlighted
region while the red highlighted area is the single phase flow regime for Echometer-Twister.
Figure 3-11 shows the pressure drop plot for Echometer-Twister and Echometer
(3X1) with respect to varying gas rates and constant liquid rates. The plot shows that the
pressure drop values for Echometer (3X1) are independent of the gas rates into the system.
Pressure drop in Echometer-Twister is independent of gas flow rates when one phase liquid
flow exists inside the gas separator. When two phase gas-liquid flow exists, increasing gas
rates increase the pressure drop across the gas separator.
71
Pressure Drop Analysis for Echometer (3X1) and Echometer-Twister at Constant Liquid Rates
12
10
E 463BPD
E 386BPD
E 277BPD
ET 644BPD
ET 560BPD
ET 434BPD
E= Echometer (3X1)
ET = EchometerTwister
0
0
20
40
60
80
100
120
140
Figure 3-12- Pressure drop for Echometer-Twister and Echometer (3X1) at constant liquid
rates; Pc = 10 13 psi
72
3.2.2
Twister and Patterson (3X1) gas separators. The performance plot in field units is given in
Figure 3-14.
Separator Type: Patterson Twister; Patterson (3X1)
OD Dip Tube = 1; Number of Slots = 16; Casing Pressure = 10 13 psi
Dimension of Slots =8 X 1/8; Position of the Separator = Above the Perforations
Patterson Twister
Patterson (3X1)
Figure 3-13 - Comparison of Patterson Twister and Patterson (3X1) results in terms of
superficial velocity
73
Patterson Twister
Patterson (3X1)
Figure 3-14 - Comparison of Echometer Twister and Echometer (3X1) results in Field
Units
74
Figure 3-13 show that there is a noticeable change in the threshold no gas
downward Vsl. It is observed that the Vsl increased from 6 inch/sec for the Patterson (3X1)
green highlighted area, to 10 inch/sec for the Patterson Twister (red highlighted area).
Figure 3-14 shows that this increase meant that Patterson Twister gas separator could
completely separate gas from liquid at liquid rates up to 475BPD for all gas rates tested and
at some certain higher liquid rates within the red highlighted area.
In the laboratory test to determine the pressure drop within the separator with single
phase liquid flow into the laboratory test well, it is seen from Figure 3-15 that Patterson
Twister gas separator yielded higher pressure drop values which were not significantly
different from pressure drop values of the Patterson (3X1) design.
Pressure Drop between entry ports and pump intake for 100% liquid and required liquid submergence
7
0
0
100
200
300
400
500
600
700
Patterson (3X1)
Figure 3-15 - Pressure Drop between the entry ports and pump intake for Patterson
Twister and Patterson (3X1) separators; Casing Pressure (Pc) = 10 13psi
75
Figure 3-16 shows that the pressure drop across the Patterson Twister gas
separator and Patterson (3X1) gas separator depend on the liquid rates.
The same phenomenon observed for the comparison of Echometer (3X1) and
Echometer-Twister in page 69 is noticed in Figure 3-16. Some pressure drop values are
lower for Patterson-Twister compared to Patterson (3X1) at constant gas rates. Like
Echometer-Twister, those values occurred for single phase liquid flow condition in the gas
separator, see Figure 3-14. Patterson (3X1) had higher pressure drop at those liquid rates
because two phase gas-liquid flow was existent in the gas separator.
Pressure Analysis for Patterson(3X1) and Patterson-Twister at constant gas rates
10
PT 13 MSCFD
PT 41 MSCFD
PT 80 MSCFD
PT 115 MSCFD
P 13MSCFD
P 39MSCFD
6
P 76MSCFD
P 110MSCFD
P = Patterson(3X1)
4
PT =
PattersonTwister
3
2
1
0
0
100
200
300
400
500
600
700
Figure 3-16 Pressure drop for Patterson-Twister and Patterson (3X1) at constant gas rates;
Pc = 10 13 psi
76
PT 554BPD
PT 482BPD
P 650BPD
P 566BPD
P 426BPD
P 276BPD
P = Patterson(3X1)
PT = PattersonTwister
3
2
1
0
0
20
40
60
80
100
120
140
Figure 3-17 - Pressure drop for Patterson-Twister and Patterson (3X1) at constant liquid
rates; Pc = 10 13 psi
Figure 3-17 shows that gas rates do not affect pressure for Patterson (3X1) gas
separator. The values are relatively constant over constant liquid rates. The pressure drop
values are linked by blue lines. The red lines link pressure drop values for Patterson-Twister.
They are constant for single phase liquid flow through the separator. The values start to
differ when two phase flow starts. For example, the 635 BPD red line shows that two phase
gas-liquid flow becomes the controlling flow regime for gas rates of 41 MSCFD and above
at that liquid rate.
77
3.2.3
constructed to study the effect the number of dip tube twists inside a static centrifugal
separators had on gas liquid separation. Figure 3-18 is a Patterson Twister gas separator
with 2 twists. The initial Patterson Twister gas separator has 4 twists.
1
2
78
79
Figure 3-20 - Patterson Twister (2 twists) gas separator results in field units
In the plots that follow a comparison is made between the initial Patterson Twister
with 4 twists and Patterson Twister 2 twists.
80
Figure 3-21 - Comparison of Patterson Twister (4 twists) and Patterson Twister (2 twists)
results in superficial velocity terms
81
Figure 3-22 - Comparison of Patterson Twister (4 twists) and Patterson Twister (2 twists)
results in Field Units
In Figure 3-21 the performance plot for the Patterson Twister 2 twist gas
separator is superimposed on that of Patterson Twister 4 twist gas separator. The red
region represents the optimum operational area for the Patterson Twister 4 twists (the
initial design). It is seen that the threshold downward Vsl for the Patterson Twister 2
twist is noticeably less than that of the Patterson Twister 4 twists gas separator. Patterson
82
5
1
4
8
3
6
2
4
0
0
100
200
300
400
500
600
700
Patterson-Twister (4 twists)
Figure 3-23 - Pressure drop between the entry ports and pump intake for Patterson
Twister 2 twists and 4 twists gas separators; Casing Pressure (Pc) = 10 -13 psi
83
3.2.4
centrifugal separators. All the gas separator designs have 4 twists and differ only in their
entry port geometry.
In Figure 3-24 it seen that the threshold downward superficial liquid velocity inside
the separator for all the static centrifugal gas separator designs was between 9.5 -10 inch/sec.
It is also observed that the separation performance at higher liquid and gas rates yielded very
similar results. This observation agrees with Guzman6 whose work studied different entry
port geometries and concluded that the entry port geometry does not control the gas
separator liquid gas separation performance as long as the total entry port area is at least
65% of the separator annular area.
In field units Figure 3-25 the threshold downward Vsl is between 475 480 BPD for
all the static centrifugal separators.
Figure 3-26 is a pump liquid fraction plot at 10 in/sec for all the static centrifugal
separators. The plot shows that the pump fill was 100% for all the designs.
84
Twister
Patterson Twister
Echometer Twister
Figure 3-24 Comparison of results for all static centrifugal separators in terms of
superficial velocities
85
Twister
Patterson Twister
Echometer Twister
Figure 3-25 - Comparison of results for all static centrifugal separators in field units
86
0.95
0.9
PattersonTwister
EchometerTwister
0.85
Twister
0.8
0.75
0.7
0
10
20
30
40
50
60
70
Figure 3-26 - Pump Liquid Fraction for Static Centrifugal Separators at 10 in/sec
87
3.3
88
Section 3.3.1 covers all the effects of reducing the casing annular area and
increasing the separator annular area for Echometer gas separator designs.
Performance plots of Echometer (4X1), Echometer (3X1), Echometer
(4X1.5), Echometer (3X1.5) and Echometer (4X1.75) will be used. Analysis
will be based on;
o Performance plots in terms of field units and superficial velocities,
o Pump liquid fraction plots with respect to superficial gas velocity in
the casing and,
o Pressure drop profiles for single phase liquid flow and two phase gasliquid flow.
The sequence of analysis listed above is repeated for Patterson gas separator
designs in Section 3.3.2. The performance plots for Patterson (3X1),
Patterson (3X1.5), Patterson (4X1), Patterson (4X1.5), Patterson (4X1.75)
and Patterson (4X2) will be used.
89
3.3.1
separator performance.
Table 2-2 presents a summary of Echometer design configurations. From the table it
seen that the cross sectional area of the casing annulus is reduced by 6 sq. in. and the
cross-sectional area of the separator annulus increases from 4.17 in2 to 10.26 in2 according to
the size of the dip tube OD.
Figure 3-27 is a comparison of Echometer (3X1) and Echometer (4X1) and Figure
3-28 is a comparison of Echometer (3X1.5) and Echometer (4X1.5). The green highlighted
regions depict the area where no gas entered into the dip tube suction of the 4 inch OD gas
separators. The red highlighted areas in the plots are no gas zones for the 3 inch OD gas
separators.
Both plots show that tests were conducted for liquid rates up to 900 BPD for the 4
inch OD separators as against 500 BPD for the 3 inch OD separators. The improvement in
the liquid handling capacity was a direct consequence of reduction in the downward liquid
superficial velocity. The reduction in downward Vsl is a result of increase in separator annular
cross- sectional area. The reduction in downward Vsl forced the 2 phase gas liquid flow into
flow regimes where gas-liquid separation is possible. Section 3.5 of this thesis deals on flow
regimes inside the separator and provides further explanatory details on the relationship
between downward Vsl and flow regimes inside the separator annulus.
90
Echometer (3X1)
Echometer (4X1)
Figure 3-27 - Echometer (4X1) and Echometer (3X1) results compared in field units
91
Echometer (3X1.5)
Echometer (4X1.5)
Figure 3-28 - Echometer (4X1.5) and Echometer (3X1.5) results compared in field units
92
Figure 3-27 and Figure 3-28 show that the optimum operating area for the 4 inch
OD separators (green highlighted area) are larger than their 3 inch OD counterparts (red
highlighted region). An increase of 125 BPD in gas free liquid production is achievable if a 4
inch OD Echometer gravity driven gas separator replaces a 3 inch OD Echometer gas
separator under similar laboratory test conditions.
This thesis does not haste to conclude that 4 inch OD Echometer gas separators
perform more efficiently than the 3 inch OD gas separators until both designs can be
comparatively compared in terms of superficial velocities. The superficial velocity is a ratio
of the liquid/gas rates to the annulus cross sectional area as shown in Equation 4
Q
Vs =
A
4
Where Q = liquid/gas rate; A = cross sectional flow area (hydraulic area); Vs= superficial
liquid/gas velocity
Figure 3-29 presents the performance results for Echometer (4X1.75) and Figure
3-30 compares all Echometer gas separators in terms of superficial velocities.
93
94
Echometer (4X1)
Echometer (4X1.5)
Echometer (4X1.75)
Echometer (3X1.5)
Echometer (3X1)
Figure 3-30 Comparison of results of all Echometer gas separators in terms of superficial
velocity
95
Figure 3-30 is a comparative analysis for both the 3 inch OD and 4 inch OD gas
separators. Reduction of the casing annular area resulted in an increase in the range of
superficial liquid velocity from 72 to 96 in/sec. The increase in separator annular area caused
a reduction of liquid superficial velocity inside the separator from 13 to 10 in/sec.
The red highlighted area is the optimum operational region for the Echometer 3 inch
OD gas separators; it has a threshold value of 7 in/sec. The threshold value for the 4 inch
OD gas separators is between 5.5 in/sec and 6in/sec which is slightly less than the threshold
value for the 3 inch OD separators. It is shown on the plot as the green highlighted area.
Analyzing all the data acquired from the plots it can be concluded that the 6in/sec industry
rule of thumb is a valid parameter for designing efficient gravity driven separators. It is also
seen that increasing or decreasing separator annular cross sectional annular area does not
affect gas separator performance so long as the liquid production rate is within the threshold
downward superficial liquid velocity for gravity driven separators.
Further analysis using the pump liquid fraction showed that all the Echometer
separators performed within similar ranges when compared at two downward liquid velocity
values, shown in Figure 3-31 and Figure 3-32. It is clearly seen that at the threshold
downward Vsl the pump fill percentage was between 92 100% and at 10in/sec pump fillage
was between 80% and 85%.
96
0.95
0.90
0.85
0.80
0.75
0.70
0
20
40
60
80
100
Echometer (3X1)
Echometer (4X1)
Echometer (4X1.75)
Echometer (3X1.5)
97
120
0.95
0.90
0.85
0.80
0.75
0.70
0
20
40
60
80
100
120
Echometer (4X1.5)
Echometer (3X1)
Echometer (4X1)
Echometer (4X1.75)
Figure 3-33 is the performance plot exclusively for the 4 inch OD Echometer gas
separators. The aim is to study the effect of reducing/increasing separator annular cross
sectional area for 4 inch OD Echometer designs.
98
Echometer (4X1)
Echometer (4X1.5)
Echometer (4X1.75)
Figure 3-33 Comparison of all Echometer 4 inch OD separator results in field units
It is seen from Figure 3-33 that Echometer (4X1.5) and Echometer (4X1.75) gas
separators had identical amounts of gas rates passing through the dip tube suction at similar
input liquid and gas rates. The gas rates through the dip tube for Echometer (4X1) gas
separator were the largest amongst the 3 Echometer 4 inch OD combinations.
Echometer (4X1) produced the worst performance even though the separator
annular area was largest amongst the 3 designs. The reason for this result has not been
clearly understood.
99
Figure 3-34 presents the pressure difference from the ports to the separator
discharge 4 inch OD Echometer gas separators for a 2 phase gas liquid flow. The plot
shows that Echometer (4X1) has the highest pressure drop across the separator at similar gas
and liquid rates. The plot also shows that the pressure drops for 4 inch OD Echometer
designs are liquid rate dependent and are independent of the annular gas rates. The blue
arrow highlights the consistency of pressure drop values along a constant liquid rate and
varying gas rate. The red arrow shows abrupt changes in pressure drop at increasing liquid
rates for a constant gas rate.
Figure 3-35 shows the pressure drop for single phase liquid flow for all Echometer
designs. The plots show that Echometer (4X1) and Echometer (3X1) designs had
distinctively higher pressure drops for all the liquid rates tested due to the small internal
diameter of the dip tube. All the other designs had very similar results.
100
Echometer (4X1)
Echometer (4X1.5)
Echometer (4X1.75)
Figure 3-34 - Pressure drop between the entry ports and pump intake for Echometer 4 inch
OD gas separators for 2 phase gas liquid flow; Casing Pressure (Pc) = 10 -13
psi
101
Pressure Drop between the entry ports and pump intake for 100% liquid and the required Liquid
Submergence for Echometer Down-hole gas separator designs
9
6
1
5
1
4
8
3
6
2
0
0
100
200
300
400
500
600
700
800
900
1000
(4X1.5)
(4X1)
(3X1)
(3X1.5)
Echometer (4X1.75_DT=4'5'')
Figure 3-35 - Pressure drop between the entry ports and pump intake all tested Echometer
gas separators; Casing Pressure (Pc) = 10 -13 psi
102
3.3.2
Table 2-3. The differences are seen in the reduction of cross sectional area of the casing
annulus. The separator annular area varies with OD of dip tube used in the gas separator.
Figure 3-36 represents the comparison of the performance plots for Patterson (3X1)
and Patterson (4X1). Figure 3-37 shows the comparison for Patterson (4X1.5) and Patterson
(3X1.5). In both plots it is seen that the liquid handling capacity for the 4 inch OD Patterson
gas separators are at least twice the maximum for the 3 inch OD types. The reason for this
increase in liquid handling capacity is associated to the reduction in downward Vsl inside the
separator because of the increase in separator annular area. The optimum operational area
for the 4 inch OD Patterson separator is larger than the 3 inch OD gas separator. A
production rate difference of 125 BPD separates the optimum production rate for the 3 inch
OD separators from the 4 inch OD separators which is shown from the plots to be
450BPD.
Figure 3-38 and Figure 3-39 show the performance plots for Patterson (4X1.75) and
Patterson (4X2). Both gas separator designs have a threshold gas free liquid production rate
of 435 BPD.
Figure 3-40 clearly illustrates the effect of increasing the interior flow area and
reducing the exterior flow area. Comparison is made based on superficial velocities.
103
Patterson (3X1)
Patterson (4X1)
Figure 3-36 - Patterson (4X1) and Patterson (3X1) results compared in field units
104
Patterson (3X1.5)
Patterson (4X1.5)
Figure 3-37- Patterson (4X1.5) and Patterson (3X1.5) results compared in field units
105
106
107
Patterson (4X1)
Patterson (4X1.5)
Patterson (4X1.75)
Patterson (4X2)
Patterson (3X1)
Patterson (3X1.5)
Figure 3-40 - Comparison of results for all Patterson 3 inch OD and 4 inch OD separators
in superficial velocity terms
108
The effect of increasing the separator annular area resulted in the reduction of
downward liquid superficial velocity from 14 in/sec to 12 in/sec. The maximum liquid rate
of 900 BPD for 4 inch OD Patterson separators shown in previous plots yields a downward
superficial liquid velocity of 12 in/sec. The 450 BPD liquid input rate for the 3 inch OD
separators is equivalent to 14 in/sec downward Vsl. The reduction in downward superficial
liquid velocity from 14 to 12 in/sec resulted in a 450 BPD increase in the volumetric rate
handled by the 4 in OD gas separator.
The reduction of casing annulus cross- sectional area increased the casing superficial
gas velocities increased by 24 in/sec.
At 6 in/sec all the separator designs separated all the gas in the 2 phase flowing
liquid. The industry rule of thumb is valid for Patterson type gravity driven separators. At the
6 in/sec threshold an increase or decrease in separator annular cross sectional area does
not affect separation performance. The exact same result was seen in Section 3.3.1, page 95,
for Echometer type separators.
Pump liquid fraction analysis is done for the Patterson type designs to further
validate the concept that all gravity driven separators yield relatively similar results at similar
downward superficial velocities. Two downward Vsl values are evaluated in Figure 3-41 and
Figure 3-42.
In Figure 3-41 all the Patterson separator designs excluding Patterson (4X1) had
pump fillage efficiencies between 92 100%. At downward liquid velocities between 8 9
in/sec all the designs had a very similar trend with efficiencies between 75% and 95%
109
0.95
0.9
0.85
0.8
0.75
0.7
0
10
20
30
40
50
60
70
80
90
Patterson (4X1.75)
Patterson(4X1.5)
Patterson (4X1)
Patterson (3X1)
Patterson (3X1.5)
110
100
0.95
0.90
0.85
0.80
0.75
0.70
0
10
20
30
40
50
60
70
80
90
Patterson (4X1.75)
Patterson(4X1.5)
Patterson (4X1)
Patterson (3X1)
Patterson (3X1.5)
Figure 3-42 - Pump Liquid Fraction for Patterson Separators between 8 9 in/sec
111
100
112
Separator: Patterson (4X1) & Patterson (4X1.5) & Patterson (4X1.75) & Patterson (4X2)
Number of Slots = 16; Dimension of Slots = 8X ; Position of Separator = Above Perforations
Test Type: Continuous Flow; Pc = 10psi 13 psi
Patterson (4X1)
Patterson (4X1.5)
Patterson (4X1.75)
Patterson (4X2)
Figure 3-43 - Comparison of all Patterson 4 inch OD separator results in field units
113
Pressure Drop between entry pots and pump intake for 100% liquid in the system for patterson
Down-hole gas separators
9
8
7
6
5
4
3
2
1
0
0
100
200
300
400
500
600
700
800
900
100
Patterson 4x1.75
Patterson 4x1.5
Patterson 4x1
Patterson 3x1
Figure 3-44 Pressure Drop between the entry ports and pump intake for Patterson 4 inch
OD gas separators; Casing Pressure (Pc) = 10 13psi
114
Figure 3-45 - Pressure drop between the entry ports and pump intake for Patterson 4 inch
OD gas separators at varying gas and liquid rates; Casing Pressure (Pc) = 10 13psi
115
3.4
To study the effect of the dip tube length in Echometer 4 inch OD gas separators
Echometer (4X1.75) dip tube was shortened by 2 ft7in. since the experimental set-up does
not allow increasing the length of the laboratory test well.
Figure 3-41 and Figure 3-42 show that the separation performance of Echometer
(4X1.75) gas separator was unhampered by shortening the dip tube from 5ft to 2ft.
116
Separator: Echometer (4X1.75) & Echometer (4X1.75) with Dip tube length 2ft
Number of Slots = 4; Dimension of Slots = 4X 2; Position of Separator = Above Perforations
Test Type: Continuous Flow; Pc = 10psi 13 psi
Figure 3-46 Comparison of results for Echometer (4X1.75) with 5dip tube and Echometer (4X1.75)
with 2 dip tube in superficial velocity terms
117
Separator: Echometer (4X1.75) & Echometer (4X1.75) with Dip tube length 2ft
Number of Slots = 4; Dimension of Slots = 4X 2; Position of Separator = Above Perforations
Test Type: Continuous Flow; Pc = 10psi 13 psi
Figure 3-47 - Comparison of results for Echometer (4X1.75) with 5dip tube and Echometer (4X1.75)
with 2 dip tube in field units
118
3.5
pumping 4(SRP) Both valves control the injection and expulsion of fluids into the pump
barrel during the down-stroke and upstroke motion of the sucker rod. During the down
stroke the traveling valve (TV) will open when the pressure below the valve exceeds the fluid
pressure above the valve. During the upstroke the standing valve will open when the well
pressure below the valve exceeds the pressure above the valve.
This experiment applies to the up - stroke action of the SRP when the standing valve
is completely open, which is when the fluid from the down hole separator enters into the
pump barrel.
The objective of the experiment was to study the effect of including a 1 standing
valve assembly (H-F-134E22 307) shown in Figure 3-48 and attached to the gas separator
head as seen in Figure 3-49.
Continuous flow tests were carried out on the new combination and the results are
presented Figure 3-50 through Figure 3-51
119
120
121
122
3.5.1
Figure 3-52 and Figure 3-53 represent the comparisons in terms of superficial
velocity and in field units.
Separator: Echometer (3X1) & Echometer (3X1) with Standing Valve joint
Number of Slots = 4; Dimension of Slots = 4X 2; Position of Separator = Above Perforations
Test Type: Continuous Flow; Pc = 10psi 13 psi
Echometer (3X1)
Echometer (3X1) with SV
Figure 3-52 Comparison of Echometer (3X1) with and without Standing Valve in terms of
superficial velocities
123
Separator: Echometer (3X1) & Echometer (3X1) with Standing Valve joint
Number of Slots = 4; Dimension of Slots = 4X 2; Position of Separator = Above Perforations
Test Type: Continuous Flow; Pc = 10psi 13 psi
Echometer (3X1)
Echometer (3X1) with SV
Figure 3-53 - Comparison of Echometer (3X1) with and without Standing Valve in field units
From Figure 3-52 and Figure 3-53 it is observed that the addition of a standing valve
between the gas separator and the tubing return line (pump intake) did not alter the
separation performance. The next section studies the effect of standing valve on the
separator pressure drop.
124
3.5.2
The pressure drop analysis is shown in Figure 3-54 for 100% liquid flowing through
the test well. Figure 3-55 shows the calculated pressure drops at representative liquid and gas
rates
Pressure Drop between Entry Ports and Pump Intake for 100% Liquid in the Laboratory Test Well
5
0
0
100
200
300
400
500
600
Figure 3-54- Pressure drop between the entry ports and pump intake for Echometer (3X1)
with and without Standing Valve
125
Echometer (3X1)
Echometer (3X1) with SV
Figure 3-55 - Pressure drop between the entry ports and pump intake for Echometer (3X1)
with and without Standing Valve at varying gas and liquid rates; Pc = 10 -13 psi
The preceding comparisons in terms of separation performance and pressure drop
for continuous flow have shown that practically there is effect due to addition of standing
valve assembly.
3.6
test of the static centrifugal separators5 (Patterson Twister and Echometer Twister);
Bohorquez had done an earlier analysis in his thesis on flow regimes of static centrifugal separators.
126
bubble flow, churn flow and annular flow. During the test of the all the 4 inch OD
separators (Echometer and Patterson designs) only bubble flow regime was observed.
Bubble flow regime occurs when liquid is the continuous phase and gas is dispersed
into bubbles. The annular flow regime takes place when the gas flows in the core if the
annular cross - sectional area and the liquid flows as a film around the gas separator and the
dip tube walls. The annular flow regime is also referred to as waterfall in this study.
Sometimes the annular flow stops at a distance before the bottom of the dip tube and the
flow becomes a turbulent mixture where none if the phases, gas or liquid are continuous;
this flow regime is referred to as churn flow6. Figure 3-56 illustrates the three types of flow
regimes.
Figure 3-56 Flow Regimes observed in the gas separator annular area (courtesy Renato
Bohorquez7)
6
127
Figure 3-57 maps the flow regimes observed for the 3 inch OD gas separators7 as
analyzed by Bohorquez7. The flow regimes are mapped for combinations for superficial gas
velocities in the casing annulus and superficial liquid velocities in the separator annulus. The
flow regimes dominating each region of the map are labeled as liquid flow, bubble flow or
waterfall. Liquid flow means that no gas was entering the dip tube suction. Figure 3-78 maps
the flow regime(s) observed for 4 inch OD gas separators tested by this author.
Figure 3-57 Flow regime map for the annular space of 3 inch OD gravity driven gas separators7
128
12
10
Bubble flow
8
6
Bubble Flow
Liquid Flow
Liquid flow
2
0
0
20
40
60
80
100
120
Figure 3-58 Flow regime map for the annular space of 4 inch OD gravity driven gas separators
From the flow regime maps Figure 3-58 and Figure 3-57, it seen that the liquid flow
threshold is at approximately 6 inch/sec downward superficial liquid velocity. It is also
observed that the bubble flow region exists at downward liquid velocities between 7 -12
inch/sec for both 3 inch and 4 inch OD gas separators.
Earlier discussions on 4 inch OD separators performance plots had shown their high
liquid handling capacity 8- almost double the 500BPD liquid handling capacity of 3 inch OD
separators. The reason for the improvement can be seen from the flow regime maps in
Figure 3-57 and Figure 3-58. 3 inch OD gas separators enter the water fall condition at Vsl =
12 in/sec and a casing Vsg = 40 in/sec a situation where it is practically impossible for gasliquid separation to occur. In the 4 inch OD separators it is seen that at the same superficial
8 Liquid handling capacity for 4 inch OD separator is expected to be higher than stated in this thesis report.
The present laboratory set up supplies a maximum liquid rate of 900 BPD of water into the liquid flow
loop.
129
liquid and gas velocities of 12 in/sec and 40 in/sec bubble flow condition exists. At this flow
condition gas-liquid separation can occur inside the separator annulus.
The plots also show that the maximum superficial liquid velocity in the separator
annulus to be noticeably higher for 3 inch OD gas separators compared to 4 inch OD gas
separators even though the liquid handling capacity for the latter is at least twice that of the
former (see dashed red lines in the plots; 14 in/sec for 3 inch OD gas separators and 11
inch/sec for 4 in OD gas separators).
The increase in separator annular area reduced the downward liquid superficial velocity
in the separator annulus.
Within the context of these experiments the reduction in downward superficial liquid
velocity forced the liquid and gas rates inside the separator annulus into flow regimes where
gas-liquid separation is possible.
Figure 3-59 show the flow regime map for the Twister analyzed by Bohorquez7 and
Figure 3-60 show the flow regime map for Patterson Twister and Echometer Twister.
130
Figure 3-59 Flow regime map for the Twister separator annlus7
131
16
Partial Waterfall
Waterfall
14
Bubble flow
12
10
Liquid flow
0
0
10
20
30
40
50
60
70
Bubble Flow
Water Fall
Figure 3-60 Flow regime for Patterson Twister and Echometer Twister static
centrifugal separators
Figure 3-59 and Figure 3-60 show similarities in the downward liquid velocity
threshold of 10 inch/sec although in Figure 3-60 the threshold appears uniform for all the
superficial gas velocities in the casing. The bubble flow region is slightly enlarged and the
partial water fall region slightly reduced. The water fall region remains same.
The red line in Figure 3-60 indicates the 6 in/sec limiting downward Vsl for 4 inch
OD gas separators in liquid flow regime. This value is less than 10 in/sec limiting Vsl for 3
inch OD static centrifugal separators. This difference in liquid flow regime makes 3 inch OD
static centrifugal gas separators more efficient than 4 inch OD gravity driven separators.
132
Chapter 4
Bubble Rise Experiments
The premise of this work is based on the bubble slip velocity. In Chapter 3 a lot of
mention was made about downward superficial velocity as a valid parameter for designing
down hole gas separators. Optimum separation efficiencies are achieved inside the gas
separator if the bubble rise velocity (slip velocity) is greater than the downward Vsl.2,3,4,5,6,7
The 6in/sec industry rule of thumb is based on gas bubble rise velocity in low
viscosity fluids. At this rate it has been proven that relatively large gas bubbles rise faster
than downward flowing two phase, gas and liquid mixtures inside the gas separator annulus
resulting in good separation efficiency. The question is what would be the separator
performance for liquids of higher viscosity. This can only be answered experimentally and is
one of the proposed extensions of this work. A review of the literature showed that there is
very little information about the effect of viscosity on bubble rise velocity in an annular
geometry. Consequently the following preliminary study was undertaken with the objective
to determine the variation of single bubble rise velocity for a broad range of viscosities and
thus bracket the possible range of applicability of the current results of separator
performance.
133
4.1
presentation of results
this enabled gas bubbles to be observed and captured through digital video means. Gas was
injected into a static liquid column through a vent 3 inches from the base of the acrylic glass
tube. The acrylic glass tube had a diameter of 2 inches and a 1 inch dip tube was inserted
into the tube to allow measurements in the annulus over a distance of 2 feet. The captured
video data was downloaded into an editing program where the velocities of various bubbles
sizes are measured. Different viscosity fluids were measured during these experiments.
134
liquid
level
Dip Tube
72
Acrylic Tube
gas bubbles
injected air
Figure 4-1 Schematic of Laboratory Constructed Apparatus for testing bubble rise velocity
Table 4-1 Dimensions of bubble rise experiment apparatus
135
4.2
4.2.1
Glycerin, glycol and corn syrup were tested to correctly prove that they were
Newtonian fluids.
Figure 4-4through Figure 4-6 show that there is a linear relationship between shear
stress and shear rate to demonstrate the Newtonian nature of the fluids.
136
137
Apart from the Newtonian nature of the fluids their water solubility was given crucial
consideration. This consideration was necessary so as to perform experiments over a wider
range of viscosities.
Water was added in controlled volumes to glycerin and corn syrup to achieve desired
viscosities. The following tables and plots provide data for volume percentages of the test
fluids and water and their corresponding viscosities at room temperature.
138
Water
%
0
8.57
12.86
14.29
20
28.57
40
0
Viscosity
cP
800
300
110
90
50
20
10
0
Figure 4-7 Viscosity plot for Glycerin in association with water at room temperature
139
Table 4-4 - Test data for corn syrup in association with water
Water (%)
0
10
15
20
25
30
50
0
Viscosity
(cp)
3500
470
250
210
70
50
20
0
Figure 4-8 Viscosity plot for Corn Syrup in association with water at room temperature
140
Figure 4-9 Combined viscosity plots for glycerin and corn syrup in association with water
at room temperature
4.3
velocities in static viscous liquid medium calculated over a range of viscosities as shown in
Figure 4-11.
141
c
a
Fig 4-10 Examples of bubble diameter sizes measured (a) inch (b) inch c 1inch
142
16
14
12
10
0
0
0.2
0.4
0.6
0.8
1.2
1.4
glycol ( = 19cp)
glycerin ( = 800cp)
DT = dip tube OD
1.6
Other observed trends match the theory from stokes law expressed in chapter one:
larger size bubbles rise faster than small size bubbles; bubble rise velocity is dependent on
viscosity the more viscous a fluid is the less the bubble rise velocity. These trends from
stokes law are evident in Figure 4-11.
The effect of decreasing the annular area with the use of 1.5 inch OD dip tube did
not affect the bubble rise velocity as seen from crude oil rise velocity plots.
144
Chapter 5
Conclusions and Recommendations
The following conclusions and recommendations are based on the study of:
5.1
CONCLUSIONS
The conclusions are divided into four sections:
both gravity driven and static centrifugal gas separators in terms of superficial
liquid and gas velocities inside the separator annulus and in the casing
annulus,
145
rise velocity of bubbles in the annulus of static viscous medium under room
temperature.
5.1.1
146
5.1.2
1. The threshold downward superficial liquid velocity for both the 3 inch and 4
inch OD gravity driven separator remained at 6 inch/sec. The industry rule
of thumb is valid.
147
2. The gas free operational area for 4 inch OD gravity driven gas separators is
55% greater than 3 inch OD gravity driven separators.
3. Only liquid flow and bubble flow regime were observed in the 4 inch OD
gravity driven gas separator for water rates up to 900BPD.This explains the
high liquid handling capacity of 4 inch OD gravity driven gas separators..
4.
5. The pressure drop within the gas separator for both 3 inch OD and 4 inch
OD gravity driven gas separators is liquid rate dependant and invariant to gas
rates.
6. Separator designs with 1 inch OD dip tube sizes yielded the largest pressure
drops within the gas separator.
7. Reducing the separation distance by shortening the dip tube length from
5ft to 2ft for the 4 inch OD gravity driven separator does not affect
separation performance.
8. Only liquid flow and bubble flow regime was observed in the 4 inch OD
gravity driven gas separator for water rates up to 900BPD.This explained the
high liquid handling capacity of 4 inch OD gravity driven gas separators.
9. The reason(s) for the poor performances of Echometer (4X1) and Patterson
(4X1) is yet to be determined.
148
5.1.3
1. The rise velocity for gas bubble diameter sizes , and 1 in a static,
vertical liquid column in an annulus with viscosities between 1 and 110 cp,
range from 10 12 in/sec.
2. Results from down hole gas separator experiments using water, so far
analyzed, that showed the optimal liquid Vsl ranges between 6 and 10 in/sec
should also be representative of the performance for fluids with viscosities
less than 110 cp..
5.2
149
4. .The inner diameter of the dip tube should be the largest possible allowed by
the casing and gas separator diameters7
5. Maximize the separator annular area by using thin wall in designing the
separator outer wall. But make sure that the gas separator walls have
sufficient strength.10
6. For a 4 inch OD gravity driven separator, a dip tube not less than 1.5 in OD
should be used.
7. Design gas separator using an outer wall diameter that results in less than an
100 in/sec gas superficial velocity as permitted by casing ID and tubing OD
measurements.
8. Smaller OD gas separators can perform as well as large OD separators only
at lower liquid rates.
9.
5.3
150
3. Perform tests with each separator with a higher viscous liquid(s) to verify or
disprove the conclusions made from the bubble rise experiment as stated in
this thesis.
4. Construct and run tests on Patterson and Echometer 4 inch OD designs,
with helical dip tubes to study the effect of increasing the annular area and
possibly increasing the number of twists in static centrifugal separators.
151
Appendix A
Schematics of the Echometer Separators
152
153
154
Appendix B
Schematics of the Patterson Separators
155
156
157
158
Appendix C
Original data files
The original data files are available on CD ROM by request to the Department of
Petroleum and Geosystems Engineering librarian.
http://www.pge.utexas.edu/reading/info.cfm
The sizes of the data files are too large to be included in this Thesis report.
159
Nomenclature
A
Area
Bg
Gas formation volume factor (volume of gas per barrel of oil scf/stb)
Bo
Bw
GOR
Re
Pressure (psig)
Pc
P1
P2
P3
Rs
Temperature (oF)
Volume (bbl)
Velocity (inch/sec)
vsg
vsl
vm
Casing
Gas
Liquid
160
Mixture
Pump
Tubing
Abbreviations
o
Celsius degrees
Fahrenheit degrees
bbl
Barrels
bbl/day
BPD
BHP
cp
centipoise
DGS
ESP
ft
feet
ft2
square feet
GLR
ID
Internal diameter
in
inches
in2
square inches
OD
Outer diameter
Kg/m3
m/s
MCFD
161
MSCFD
mm
Millimeters
PCP
psi
psia
psi absolute
psig
psi guage
PVC
Polyvinyl chloride
SCFD
sec
Seconds
spm
SRP
stb
Standard barrel
162
References
1. Schome W.P: Bottom hole Gas Separators Increase Production, Drill.and Prod.
Prac., API 1958
2. Clegg, J.D. Another Look at Gas Anchors. Southwestern Petroleum Short Course:
1989, 293 307.
3. McCoy, J.N., Podio, A.L., Woods, M.D. and Nygaard, H.S.: Field and laboratory
testing of a Decentralized Continuous Flow Gas Anchor, Presented at the 46th
Annual Technical Meeting of the Petroleum Society of CIM, May 14 17, 1995.
4. Campbell J.H and Brimhall R.M.: An Engineering Approach to Gas Anchor Design,
SPE 18826, Presented at the 1989 SPE Production and Operational Symposium
held in Oklahoma City, Oklahoma 13 14 March.
5. Lisigurski, O. The effect of geometry on the Efficiency of Downhole Gas Separators.
The University of Texas at Austin, Austin. (2004)
6. Guzman, M.. Downhole Gas Separator Performance in Sucker Rod Pumping System.
The University of Texas at Austin. (2005)
7. Bohorquez, R. Performance of Gravity Driven and Centrifugal Down hole Gas
Separators for Continuous and Intermittent Flow. The University of Texas at
Austin. (2006)
8. Kobylinski, L.S., Taylor F.T. and Brienan, J.W.: Development and Field Test Results
of an Efficient Downhole Centrifugal Separator. Journal of Petroleum
Technology July 1985, 1295 1304.
9. McCoy, J.N., Podio, A.L., Lisigurski, O., Patterson, J. and Rowlan, L.: A Laboratory
Study With Field Data of Downhole Gas Separators SPE 96619, Presented at
SPE Annual Technical Conference and Exhibition, Dallas, 9 12 October 2005.
10. McCoy, J.N., Podio, A.L.: Improved Downhole Gas Separator, SWPSC 1998011,
Presented at the 45th Annual Southwestern Petroleum Short Course.
11. Patterson, J.C., and Leonard, L.: Gas Anchor Design Changes Used to Improve Gas
Separation in Coalbed Methane Operations in Wyoming, SWPSC 2003011,
Presented at the 2003 Annual Southwestern Petroleum Short Course.
12. US 3128719, 1964-04, Jongbloed, L.J.S., and Vos, E.A.: Gas Anchor.
163
13. US 5902378, 1999-05-11, Obrejanu, M.: Continuous Flow Downhole Gas Separator
for Progressing Cavity Pumps.
14. US 3887342, 1975-06-03, Bunnelle, P.R.: Liquid Gas Separator Unit.
15. US 4981175, 1991-01-01, Powers, M.L.: Recirculating Gas Separator for Electric
Submersible Pumps.
16. US 6382317, Cobb, E. D.: Apparatus and Method for Separating Gas and Solids
from Well Fluids.
17 Bohorquez, R., Ananaba, O.V., Podio, A.L., Lisigurski, O. and Guzman, M.:
Laboratory Testing of Downhole Gas Separators, SPE 109532, Presented at the
SPE Annual Technical Conference and Exhibition, Anaheim, 11 14 November
2007.
18. Dottore, E.J., How to Prevent Gas Lock in Sucker Rod Pumps, SPE 27010,
Presented at the III Latin American / Caribbean Petroleum Engineering
Conference held in Buenos Aries, Argentina, 13 14 March 1994.
164
Vita
Ochiagha Victor Ananaba was born in Aba, Nigeria on July 31st 1980, the first
child of Col. Emeka and Lolo Nnenne Ananaba. In June 1998, he graduated from high
school at Federal Government College Ikot-Ekpene. He was granted admission into Federal
University of Technology Owerri, September 1998, where he earned a Bachelor of
Engineering degree in Petroleum Engineering in December 2003. He was drafted into the
National Youth Service Corps in August 2004 and successfully completed his national
assignment August 2005. In January 2006 he entered The Graduate School of The
University of Texas at Austin to pursue his Masters degree in Petroleum Engineering.
He will be working fulltime as an engineer for the Artificial Lift team of
Schlumberger DCS@ccess Reservoir Well Completions and Productivity group starting
February 2008.
Permanent address:
165