Sei sulla pagina 1di 15

SPE-170039-MS

ESP Optimization in an Extra Heavy Oilfield: A Case Study in Colombian


Llanos Basin
Uribe J.E. Lopez and Morales G.E. Jiménez, SPE

Copyright 2014, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Heavy Oil Conference-Canada held in Alberta, Canada, 10 –12 June 2014.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
The experience of working with extra-heavy oil producing wells profitably requires the use of available
technologies to the limits of their capabilities. Due to its heavy use of electricity, the Electrical
Submersible Pump (ESP) has often proven unsuitable for fields producing high-viscosity fluids. However,
in some cases, given the conditions downhole and fluid behavior, along with technological changes, it has
been possible to use ESPs at moderate cost.
Exploitation today has become technically and economically feasible for reservoirs in upper or lower
layers that were previously uneconomical given lifting, transportation and marketing costs.
Previously discovered resevoirs in the Llanos Basin of Colombia were considered of no commercial
value given the characteristics of its oil (higher viscosities to 1,000 cps at 60° F /14.7 psia and API
gravities less than 10°). The Operating Companies of the same blocks discovered other, more easily
exploited oil reservoir units. These oilfields, now in their mature stages, have high production costs
because of the inevitable increase in water cut and/or decreased pressure. This has spurred to the use of
existing infrastructure and advances in technology for lifting, processing and transportation of heavy and
extra-heavy oil reservoir units already discovered.
There have been several challenges in developing extra-heavy oil wells, such as organic precipitation,
presence of scales, strong emulsions and high GOR.
ESP technology has advanced in the extraction of large fluid volumes, smaller stage diameters,
metallurgy resistant to corrosion and abrasive solids, stronger shafts, and conditions of high pressure and
high temperature. However, there are very few developments in sustainable extraction of heavy oil and
extra-heavy oil.
This article describes an example of implementing the Electro Submersible Pump as a feasible method
for artificial lift in wells producing extra-heavy oil.
Introduction
The Llanos Basin is located in eastern Colombia. It covers an area of approximately 87,105 square miles
and contains potential reserves of heavy oil, estimated to be 1.8940 x1010 barrels of oil in place (OIP). The
Chichimene Field, discovered in 1969, located in central Colombia, is part of the Llanos Basin; it covers
19.17 square miles, with an OIP estimated in 2.5 x 108 barrels of extra-heavy oil. (Fig. 1)
2 SPE-170039-MS

Figure 1—Location of the Chichimene Field, Colombia.

Table 1—Reservoir Properties of Chichimene Field


Property / Unit Guadalupe K1/K2 San Fernando - T2

Depth (ft) TVDss 7000 6000


Pay Zone thickness (ft) 203 - 528 28 - 251
Reservoir pressure (psi) 3200 2600
Bubble point pressure (psi) 116 600
Salinity (ppm) 300 200 - 1200
Reservoir temperature, Tr (°F) 185 180
Oil gravity (°ApI) 16 - 20 7 - 10
Solution gas/oil ratio (scf/STB) 0 - 500 300 - 30000
Oil formation volume factor (bbl/STB) 1.0168 1.0695
Viscosity at Tr (cp) 62 1300
Porosity (%) 10 - 21% 8 - 23%
Permeability (milidarcy) 300 - 2200 30 - 12000
Initial water saturation (%) 24% 18%

In the second half of the 20th century, the Colombian llanos were explored and several reservoir finds
were made, with several heavy oil discoveres at depths of less than 5,000 ft TVD. Rod Pump, PCP and
ESP conventional systems did the lifting, with the main challenge being the transport of the fluids to the
surface dehydration facilities.
In the Chichimene Field two main producing intervals were identified: One called San Fernando
(Tertiary 2 – T2) and the other called Guadalupe (Cretaceous 1 - Chipaque K1 / 2 Cretaceous - Une K2).
The San Fernando Interval (T2) produces 7-10° API oil with viscosities ranging from 2,000-7,000 cP @
150 ° F and 800-1,500 cP @ 180 ° F and water cuts between 1% and 40% with Gas in Solution with as
reservoir drive mechanism. And Guadalupe interval (K1/K2) produces 18 ° API oil with viscosities
ranging from 150-500 cP @ 150° F and 500-800 cP @ 180° F with an active aquifer as the reservoir drive
mechanism. The properties of the intervals are in Table 1.
Background
The Chichimene Field started production in Guadalupe (K1/K2) in 1984 using Sucker Rod Pumping to
achieve an output of 5,000 bbls/d from 20 active wells. As oil production declined, because of high water
cuts, most wells were switched to ESP. Because of rising oil prices and increased demand for heavy oil,
SPE-170039-MS 3

the first well in the San Fernando (T2) interval


Table 2—Basic Data of Typical Well in the Chichimene Field
Properties Value began producing in 2005, as part of reserve devel-
Reservoir pressure (psi) 2700 opment. In December 2012, it was producing
Bubble point pressure (psi) 600 around 50,000 bbls/d with over 170 wells, of which
Reservoir temperature, Tr (°F) 180 140 wells producing formation San Fernando.
Oil gravity (°ApI) 8.0 The extra-heavy oil that is currently being pro-
Surface GOR (scf/bbl) 20000
Water specific gravity 1.05
duced in the area has gravity values ranging be-
Gas specific gravity 0.77 tween 7° and 10° API, with an average value of 8.4°
Productivity index (BFPD/psi) 0.6 API, and viscosities of 2,000 –7,000 cp at a reser-
Maximum production rate (BPD) 600 voir temperature of 150°F.
Desired total daily production (BPD) 500
To be able to exploit the extensive extra-heavy
Water cut (%) 7
Tubing head pressure (psi) 200
oil reserves economically, the drilling and produc-
Casing head pressure (psi) 100 tion technology implementations have been signif-
Casing outside diameter (in.) 9 5/8 icant over the last 20 years, particularly the use of
Tubing outside diameter (in.) 5½ horizontal wells and artificial lift systems resulting
Perforation depth (ft) well 7700 - 8500
Maximum dogleg severity (°/100 ft) 2
in new production targets of 1,000-2,000 bbls/d per
well compared to previous production of 200 bbls/d.
In view of the decline in oil production in the
Guadalupe K1/K2 interval, in 2005 perforations
were made in the San Fernando T2 interval in several wells, which yielded between 210 and 500 bbls/d
with 9° API and between 1% and 40% of water cut. From these initial tests, along with data from other
wells, crude reserves with great commercial potential were estimated. Since then, Ecopetrol has started
developing and testing existing wells in the San Fernando interval (T2) to determine the reserves and
economic feasibility.
The first wells were put into production using Sucker Rod Pumping. Runlife did not exceed 180 days
and operations often involved fishing rod string and the assembly line. Rod breakage and fluid pound, due
to the enormous frictional forces generated by the fluid on the pipe and rods, were among the many
failures. At times, the torque required surpassed 912,000 lb-in and the load on the drive was close to
42,700 lb-f, exceeding the specifications of SRP in the field. And with the limited short-term availability
of other lifting mechanism, it was decided to install ESP.
The behavior of ESP wells in the Chichimene field was thus: for producing wells of the Guadalupe
(K1/K2) interval, the MTBF was 1,000 days with a Runlife of 750 days; and for the producing wells of
the San Fernando (T2) interval, the MTBF was 540 days with a Runlife 325 days. Why this drastic
decline?
It was found that the extra heavy crude from the San Fernando (T2) interval was highly viscous oil with
precipitation of asphaltenes and water-precipitated inorganic scales. These phenomenons are sensitive to
temperature changes. Additionally, during the extensive testing, it was found that the reservoir drive
mechanism was not an active aquifer but gas in-solution, which further complicated the diagnosis of
failure and the design of the ESP systems.
Proposal for Design, Selection and Optimization for ESP for Extra Heavy
Oil
When wells produce crude oil medium density (API gravities ⬎ 20°, not susceptible to emulsions), an
important consideration in selecting the pump and the number of stages is the reservoir pressure and more
importantly, the increase in water cut during the expected Runlife of the equipment. An increase in the
water cut changes in direct proportion to the Total Dynamic Head (TDH).
In extra heavy oil with low water cuts, this premise does not seem be sufficient to consider an ESP
design. The reason is that in several cases, the Electro Submersible Pumps have been poor results using
4 SPE-170039-MS

energy loss calculations considering only water cut. If we consider a flow through vertical pipes, an
increase in viscosity has the same effect as a reduction the draw down generated by an increase in the
water cut
The concept of viscosity becomes more important as “heavy oil viscosity can change by orders of
magnitude with respect to temperature variations” (Ehlig-Economides, 2000; Trebolle, 1993). In fact,
heavy and extra heavy oil variations in viscosity caused by a decrease in fluid temperature can cause
energy losses similar to or greater than the fluid density changes caused by increases in the water cut.
The viscosity of the oil varies depending on the conditions of the well and the amount of gas
in-solution. Depending on the pressure, there are three types of viscosity. Oil viscosity at atmospheric
temperature when it contains gas in-solution, is known as “dead oil viscosity.” When the oil is saturated
with gas, is known as saturated oil viscosity or “saturated viscosity”. When the oil is at a pressure above
the bubble point (P ⬎ Pb), the oil viscosity is considered at a low level of saturation or “undersaturated
viscosity.”
High viscosity fluid can affect the ESP in different ways:
● High viscosity also reduces the pump’s ability to efficiently lift the fluid: The dynamic head
generated by each stage is reduced, requiring more stages in the pump to generate the same
pressure differential.
● The resistance to the flow increases the pump’s brake horsepower (BHP): The power consumed
per stage is increased because of increasing power requirements.
● As a consequence of a reduction of dynamic height and an increase in power needed, pump
efficiency decreases considerably.
Viscous fluid produces more friction loss in the tubing, which causes the pump to work much
harder.

Choosing ESP for extra heavy oil: Proposed changes in methodology


These are based on the usual steps used for the selection of ESP equipment in which viscosity calculations
are included:
1. Collection of Available Data
a. Dead oil viscosity at different temperatures
2. Changes in Total Dynamic Head required (TDH)
a. Build a Viscosity vs Temperature correlation to determinate an appropriate viscosity value
b. Calculations of Pressure Loss to Viscosity
3. Changes in Pump Type Selection
a. Determinate the additional power required for the pump shaft
4. Motor and Seal Selection
a. Determinate the additional power required to operate the motor

Viscosity Correlation function of Temperature for Chichimene Field


Over 100 samples corresponding to more than 60 producing wells in the San Fernando interval (T2) were
taken. These samples were subjected to rheological analysis with at least two temperature points. For the
case study, the base points were 30° C (86° F), 51° C (124° F), 65.6° C (150° F) and 82.2° C (180° F).
At first, it was determined that a polynomial correlation was most appropriate to interpolate the fluid
viscosity from the data collected. (Fig. 2)
SPE-170039-MS 5

Figure 2—Polinomial Viscosity Trend of San Fernando (T2) Oil

For samples analyzed, the initial correlation was:

Where:

␮: Dead Oil Viscosity, Centipoises


v: Cinematic Oil Viscosity, Centistokes
sg: Specific Gravity, Dimensionless
T: Temperature, °F

For the range between 90 ° F and 180 ° F correlation polynomial has a lower error rate, which is
between 2 and 6%. However, for the range of interest (between 130 and 180 ° F), the polynomial
correlation has inflection points that might be difficult to demonstrate as the viscosity increases with
temperature. Because of this the ASTM D341 method was used to construct the curve of viscosity versus
temperature. (Fig. 3)

For samples analyzed, the corrected correlation was:

Where:
6 SPE-170039-MS

Figure 3—ASTM D341 Viscosity Trend of San Fernando (T2) Oil

␮: Dead Oil Viscosity, Centipoises


v: Cinematic Oil Viscosity, Centistokes
sg: Specific Gravity, Dimensionless
T: Temperature, °F

With a model of viscosity versus temperature, we can achieve a better approximation of the behavior
of oil along both vertical and horizontal pipe and theoretical energy needs for inflow and/or pumping.
Total Dynamic Head (TDH) proposal
Consequently, the proposed Total Dynamic Head (TDH) would be:

Where:

TDH: Total Dynamic Head, feet


Hd: Vertical Distance between wellhead and fluid level, feet
Ft: Friction loss in pipe, feet. Is determined using the Hazen Williams
formula or figures.
Pd: Equivalent feet of the pressure required to overcome the pressure in the
flow line (desired head pressure equivalent feet well fluid column
height)
⌬H␮: Additional pressure of column fluid caused by high viscosity fluid, feet
SPE-170039-MS 7

Vertical Distance between wellhead and fluid level, Hd:

Where:

Hd: Vertical Distance between wellhead and fluid level, feet


SGL: Fluid Specific Gravity, dimensionless
PIP: Pump Intake Pressure (psia)

Pipe friction
The friction in the pipe will vary with the size of the pipe, its capacity, length and fluid viscosity. Tables
for calculating the friction through a piping system are available from the Hydraulic Institute Standards,
pump manufacturer’s literature and some manuals. Below is the Hazen-Williams Correlation for calcu-
lating the friction loss as a function of the pipe flow and the pipe diameter:
Hazen-Williams Correlation

Where:

⌬Pf: Pressure loss due to friction, psi per ft of pipe length


Q: Flow rate, gal/min
D: Pipe inside diameter, in
B: Hazen–Willliams roughness coefficient factor, dimensionless
B: Friction coefficient
⫽ 100 for old tubing (more than 10 years)
⫽ 120 for new tubing (less than 10 years)
⫽ 130 for fiberglass lined tubing
⫽ 140 for plastic lined tubing
L: Measured pump depth, ft

Then, this pressure differential becomes fluid column, in feet


Where:

Welhead Pressure Equivalent in feet, Pd:

Ft : Friction Loss of Pipe, feet


⌬Pf: Pressure loss due to friction, psi per ft of pipe length
SGL: Fluid Specific Gravity, dimensionless
8 SPE-170039-MS

Where:

Pd: Equivalent Fluid Column to Head Pressure, feet


THP: Tubing Head Pressure, psi
SGL: Fluid Specific Gravity, dimensionless

Inclusion of Poiseuille’s Law to calculate additional pressure loss by


viscosity
Although Poiseuille’s Law is used to calculate pressure losses in horizontal flow, for the particular case
of the Chichimene field, it is possible to apply it as an additional component of vertical flow restriction
independent of the gravitational force. This value becomes negligible when the viscosity of the fluid is
close to that of water.
Poiseuille’s Law
For laminar flow, i.e., when the Reynolds number ⬍ 2100. This function written in practical units:

Where:

Q: Flow Rate, bbl/d


␮: Fluid Viscosity, cp
L: Pipe Length, feet
D: Pipe diameter, inches
⌬P␮: Differential Pressure, psi

Where:

⌬H␮: Additional pressure of column fluid caused by high viscosity fluid, feet.

Pump Type Selection


According to standard API 11S2, the curves are based on fresh water at 60°F (S.G. ⫽1.0), give a
recommended operating range for the pump, and are commonly available for both 50 Hertz (Hz) and 60
Hertz (Hz) operation. However, the performance curve of the pump will vary significantly with the
viscosity. We have included viscosity corrections.

Viscosity Corrections
For high viscosity fluids, a flow capacity correction factor is applied to the pump performance. A graph
of available pump capacity (expressed as a percentage) vs. fluid viscosity (cP) is provided in Fig. 4.
Some authors use an equation derived from Fig. 5 to calculate the capacity correction factor. The
correction factor was applied to the pump data found in the pump catalog to determine the reduced
capacity.
SPE-170039-MS 9

Figure 4 —Pump Correction Factor for Viscous Liquids

Equation:

The range of application for the capacity reduction formula is as follows: 93 cP ⬍ ␮ ⬍ 30000 cP
An example of a pump’s performance curve at 6,200 bbls/d at different viscosities is shown in Fig. 5.

CQ: Capacity correction factor


␮: Viscosity, centipoise (cP)
A: 100.81
B: -0.00918886
C: 3.800408E-07
D: -6.13588E-12

Case Study: Description and Application


16 runs in 5 representative wells were selected: CH-A, CH-B, CH-C, CH-D and CH-E producing in the
San Fernando interval (T2). As described above, oversized ESPs were installed and their performance
evaluated. Table 3-A shows the production information for the study wells and Table 3-B shows the main
parameters of the ESPs.
10 SPE-170039-MS

Table 3-A—Production Information for Study Wells, Chichimene Field

Table 3-B—ESP Information for Study Wells, Chichimene Field

Challenges discovered
Precipitation of inorganic scales
Fluids from CH-A, CH-B, CH-C, and CH-D precipitated inorganic scales despite their low water cut.
Different scale inhibitors were tested until a substance that prevented unexpected behavior due to the
behavior of the installed motors operating at temperatures near 300° F was developed.
Gas production
Although gas volumes are moderate, it was decided to run handlers and gas separators as part of the ESP.
SPE-170039-MS 11

Figure 5—Electrical Submersible Pump Performance Curve to different fluid viscosities

Asphaltene precipitation
All wells showed potential for precipitation of paraffins and asphaltenes with values near 17%. An
asphaltene control treatment was performed and downhole dilution was postponed until a better under-
standing of the behavior of these crudes under different pressures and varying temperatures of the
reservoir and ESP intake could be achieved.
Unstable production
The flow rates of CH-A, CH-B, CH-C, CH-D were unstable, especially CH-08, so we had to collect more
data. Several multiphase flow measurement technologies were tested against estimated extra-heavy crude
flow assessment results.
From Table 3-B data, the following variables were calculated:
Nominal Pump Efficiency
Obtained from Pump Performance of each Curve.
Real Pump Efficiency
Obtained from the following correlation:

Where:

Q: Flow Rate, bbl/d


TDH: Total Dynamic Head, feet
BHP: Break Horse Power, HP
12 SPE-170039-MS

Figure 6a—Variable Frequency Pump Performance Curve for Pump 6200 bpd 1 stage, ␮ ⴝ 1 cP (CH-C)

Figure 6b—Variable Frequency Pump Performance Curve for Pump 6200 bpd 1 stage, ␮ ⴝ 2500 cP (CH-C)
SPE-170039-MS 13

Figure 7—Stages of ESP run # 1 in CH-C. Above, impellers impregnated with high hardness solids (inorganic scale). Below, break zone with plastic
deformation of the wedge

Power consumption of the ESP


Obtained from the following formula:
14 SPE-170039-MS

Where:

Pc: Power consumption, Kwh/bbl x 1000 ft


Q: Flow Rate, bbl/d
Kwh: Power consumption, Kwh
Pumpdepth: Pump Vertical Depth, feet

Changes in Variable Frequency Pump Performance Curve due to changes


in viscosity
In the ESP run number 1 for CH-C, an ESP with 6,200 bbls capacity for extracting a flow rate of 810
bbls/d with a viscosity of 1 cP, experienced downthrust. (Fig. 6a) However, thanks to the reduction in the
pump efficiency caused by the increase in fluid viscosity as indicated above (Fig 5), to extract 810 bbls/d
of a fluid with viscosity of 2500 cP, a ESP with pump capacity of 6200 bbls/d would be in range for CH-C.
(Fig. 6c)

Optimization Actions Performed


● To maintain and increase production of ESP wells the following actions were
carried out:
● Running the ESPs deeper and installing gas handles and separators which were
operationally possible, especially in wells with high contribution of gas (GOR⬎ 100
scf / stb)
● Installation of ESPs larger diameter and larger motors.
● Test ESPs floating construction mixed flow stages.
● Injection of diluent in pipelines of the gathering system in order to reduce pressure
at the wellhead.
● Design, implementation and evaluation of pilot testing of downhole injection to
reduce losses fricion around and inside the ESPs.
● Field evaluation of different Drag Reducing Agents (DRA) in both surface and
bottom in order to reduce pressure in the pipes or decrease power consumption of
the ESPs.

Results
Because gas production wells, gas handlers and gas separators were installed in the ESP, the effects of the
gas in the fluid viscosity were mitigated.
The ESP 400 series had the highest power consumption (expressed in Kw/bbl/d per 1000 ft) and lower
system efficiencies in respect to nominal values. Average consumption of around 0.09 kWh / (bbl * 1000
ft) versus 0.05 kWh / (bbl * 1000 ft) average (Up to 80% of increased consumption compared to a 500
series pump). Average efficiencies 6.41% with a minimum of 5.34% and a maximum of 7.95% (Up to
60% loss of efficiency compared to a 500 series pump).
Pumps of nominal diameter greater than 5 inches (series 513, 538, 540) were more efficient than pumps
of nominal diameter 4 inches (series 400, 456). Average efficiencies of 16.52% with a minimum of 7.65%
and a maximum of 32.45% (Up to 60% of higher efficiency compared with a 400 series pump).
SPE-170039-MS 15

The runlife of the ESP under study was between 400 and 500 days on average. Chemical injection
failures downhole allowed precipitation of paraffins, asphaltenes and inorganic scales in stages, decreas-
ing the ESP’s runlife. (Fig. 7)
Conclusions
● The behavior of the viscosity vs temperature of extraheavy oil of Chichimene field
showed a polynomial trend.
● A fluid’s high viscosity decreases the pumping capacity of an ESP.
● The viscosity of a fluid reduces the dynamic head generated by an ESP.
● For the production of extra-heavy oil, the ESP efficiency increases with the
diameter.
● Although little convincing evidence was found, a floating type pump configuration
is recommended instead of compression type pump configuration because the load
is distributed along the stages, decreasing the number of failures by broken shaft.
Acknowledgements
The authors would like to thank the Society of Petroleum Engineers for the space and opportunity to
present and publish it.

References
1. Ehlig-Economides C. et al 2000. Global Experiences and Practice for Cold Production of
Moderate and Heavy Oil. Paper SPE 58773, presented at the SPE International Symposium on
Formation Damage Control, Lafayette, Louisiana, USA, 23–24 February.
2. Ramos, et alet al: “Producing Extra-Heavy Oil From the Orinoco Belt, Cerro Negro Area,
Venezuela, Using Bottomdrive Progressive Cavity Pumps” paper SPE 97889, presented at the
2005 International Thermal Operations and Heavy Oil Symposium held in Calgary, Alberta,
Canada, 1–3 November.
3. Trebolle R. et al 1993. The Orinoco Heavy-Oil Belt Pilot Projects and Development Strategy.
Paper SPE 25798, presented at the SPE International Thermal Operations Symposium, Bakers-
field, California, USA, 8-10 February.
4. Brown Kermit E. et al 1980. The Technology of Artificial Lift Methods Volume 2b. The University
of Tulsa. Petroleum Publishing Co. Tulsa.
5. Engineering Handbook AODD Pumps, YAMADA AMERICA.
6. HI 9.6.7. Effects of Liquid Viscosity on Rotodynamic (Centrifugal and Vertical) Pump Perfor-
mance.
7. Baker Hughes Centrilift Submersible Pump Handbook 9th Edition. 2009. Claremore, Oklahoma.
BAKER HUGHES.
8. Ramos, M. and Rojas, C.: “Experiences in the Use of ESP’s in Orinoco Belt Cerro Negro Area,
Venezuela,” paper SPE 69432, presented at the 2001 Latin American and Caribbean Petroleum
Engineering Conference, Buenos Aires, Argentina, 25-28 March.
9. Gulf Professional Publishing. 2009. Electrical Submersible Pumps Manual. Design, Operations,
and Maintenance. Elsevier Inc. Burlington. USA.
10. ASTM D341-09. Standard Practice for Viscosity-Temperature Charts for Liquid Petroleum
Products

Potrebbero piacerti anche