SPE-181214-MS

Sucker Rod Pumping in the Eagle Ford Shale Field Study

Fred Clarke and Leslie Malone, Murphy Oil Inc.

Copyright 2016, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE North American Artificial Lift Conference and Exhibition held in The Woodlands, Texas, USA, 25-27 October 2016.

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
This paper presents a study of Sucker Rod Pump (SRP) systems operating in Murphy Oil's Eagle Ford
unconventional shale wells located in South Texas, USA. Lessons learned and recommended practices will
be shared which were developed to improve SRP system performance and reduce system failure frequency.
The company installed its first SRP during 2012. This system incorporated a variable speed drive (VSD),
portable generator and electric prime mover. Murphy currently produces more than 550 wells using SRP
systems across the Eagle Ford over a widely dispersed geographic region.
As the number of SRP installations grew, early life and repeat failures started to occur. The most frequent
types of failure are discussed in this paper. Lessons learned and recommended practices are shared. A study
of the entire SRP was conducted which included the surface and downhole equipment, variable speed drive
(VSD) and operating strategy. A data base was created to document information on every installation and
failure. Equipment sub-systems were created to help with analysis and included the tubing, sucker rods and
bottomhole pump.
A new set of guidelines and strategies have been developed to enhance overall system reliability and
operation. The failure data base has proven critical to identify performance gaps and specific opportunities
to improve the SRP system. The data base revealed 52% of all failures are related to the pump system, 22%
are tubing system and 26% are sucker rod system. The failure frequency is now visible for each Murphy
operating area and sub-system which helps further identify opportunities for reliability improvement. These
efforts have resulted in a failure frequency reduction of 50% from 2014 to 2016.
A number of challenges have been addressed to achieve successful SRP operation in Murphy's Eagle Ford
shale wells. Following traditional best practices and "rules of thumb" have not always delivered expected
performance and run life. This paper discusses key factors identified during the continuous improvement
journey and provides operators a benchmark of SRP system in a similar application.

Introduction
During 2009 Murphy acquired rights to 148,000 acres in South Texas in the Eagle Ford Shale. The Eagle
Ford shale is approximately 50 miles wide by 400 miles in length. First production began in 2009 in the
liquid rich portion of the field. The reservoir is accessed by drilling a horizontal well which is completed
using a fracture stimulation.
2 SPE-181214-MS

During phase 1 of the life cycle (Fig.1) the well will flow from several months to several years. The Eagle
Ford decline curve is steep and will vary dependent on lateral length, reservoir quality and connectivity.

Figure 1—Example of unconventional decline curve

There is no known natural reservoir pressure support unless there is connectivity to other zones such as
the Austin Chalk which overlays the Eagle Ford. (Martin et al. 2011)
Phase 2 of the life cycle is characterized by periods of intermittent flow. During this time, the well may
load up with dead liquid and stop producing. The well may be shut in to build up reservoir pressure. The
well may also be swabbed to unload fluid to reduce the hydrostatic head.
Phase 3 begins after energy from fracturing has dissipated and the remaining reservoir pressure is
insufficient to lift fluid to surface. Wells which appear to be incapable of flow may kick off unexpectedly
during early pump operation and sand laying in the lateral may become mobile. Critical velocity to carry
sand is easily achieved if a well flows while artificial lift is operating. In these conditions the SRP is shut
down by the operator until flow stops.
The best time to install artificial lift is when the well is incapable of flow. Installing an SRP system in a
well still capable of flow risks a premature bottomhole pump failure.
An evaluation was completed and a criteria developed based on a field development plan. The business
identified the following key attributes and requirements:

• Scalability

• Rapid deployment

• Availability in South Texas

• Ability to automate, remote control and monitor

• Minimal infrastructure

• Operate or convert from generator to utility power

• Ability to match decline curve

• Single artificial lift strategy for life of well

• Robust and simple to operate

A summary of the artificial lift system advantages and disadvantages Table 1 shows no particular
artificial lift system is perfect in all conditions. Some artificial lift systems are dependent on having a large
experienced work force and may require extensive technical support and infrastructure.
SPE-181214-MS 3

Table 1—Artificial lift comparison table

Artificial lift type Advantages Disadvantages

Sucker rod pump • Common for pumping oil wells • Low efficiency with high GLR fluid
(SRP)
• Simple operation • Difficult to draw fluid level below bubblepoint

• Wide turn down range 1.5 -7 SPM • Tubing wear operating past kick off point

• Wide range of downhole conditions • Hole in tubing if sucker rods unguided

• Automated optimization • Cannot land pump in DLS>2°

• Minimal maintenance/operators • Paraffin problem without treatment

• Produce well late life to10 bfpd • Chemical program, paraffin and corrosion

• Portable generator or utility power

• Simple integration with SCADA

Gas lift (GL) • Wire line retrievable • High pressure compression required

• Excellent gas handling • Infrastructure best to install during construction phase of

field development
• Handles solids
• High net positive suction head required
• Operation in deviated wellbores
• Methanol injection
• Compression remote from wellsite
• Freezing and hydrate on injection gas lines
• Able to convert from continuous injection to gas
assisted plunger lift • Corrosive injection gas requires treatment

• Experienced operators required

• Back pressure may decrease production

• Paraffin can be significant issue

• Decline in bottom hole pressure and productivity

• Low efficiency when producing a single well or small field

Plunger lift (PL) • No external energy required • Requires 400 scf/bbl. per 1,000 ft

• Rigless intervention • Solids can stick plunger

• Low maintenance • Requires differential pressure between bottom hole pressure

and tubing pressure to initiate plunger travel
• Tubing cleaned of paraffin and scale when plunger
travels
• Suitable for high GLR

Jet pump (JET) • Excellent gas handling • High pressure power fluid required

• Suitable for directional wellbores • High opex to maintain surface equipment

• Flexible production rate • Frequent down time associated with surface and subsurface
equipment
• Solids
• Difficult to treat for scale

• Operator required at start-up

• Large generator required and power quality

Electric submersible • No sucker rods and guides • Installation cost

pump (ESP)
• High turn down rate • Low reliability-high bottom hole temperature

• Fully automated operation • Smooth tangent section in wellbore

• Electric motor cooling

• Expensive workover to repair

• Solids

Progressing cavity • Wide turn down range • Elastomers incompatible with fluid
pump (PCP)
• Low cost • Operating temperature above capability

• Automation • Tubing and rod wear

• Low Profile • Operating depth limitation

Murphy was new to onshore operations and eager to develop the Eagle Ford asset therefore was looking
for an artificial lift system which was able scalable during early stages of field development.
4 SPE-181214-MS

Most production came from flowing wells which have a small foot print and minimal cycle time from
drill, complete to produce. Murphy's land position is widely dispersed across the Eagle Ford. This drove
the need to be able to install artificial lift one well at a time as required to maintain production.
The business selected sucker rod pumps because they are simple to install, easy to operate and common
to oil production throughout the world.

Application
Murphy operates multiple fields across the Eagle Ford shale and for the purpose of this paper wells are
grouped into three main fields. SRP systems operate a wide range of applications outlined in Table 2. The
pumps are landed in the vertical section of a well from 6,000 to 11,400 ft. Lateral lengths varying from
2,500 to 9,000 ft.

Table 2—Well profile

Field A Field B Field C

7,300 to 11,400 ft vertical depth 6,800 to 10,900 ft vertical depth 6,000 to 8,000 ft vertical depth

5.50 in. casing 5.50 in. casing 4.50 and 5.50 in. casing

2.875 in. tubing 2.875 in. tubing 2.875 and 2.375 in. tubing

Power, generator and utility Power, generator and utility Power, generator and utility

640-365-168 Pumping unit 640-365-168 Pumping unit 640-365-168 Pumping unit

456-305-144 Pumping unit

50 hp electric motor 50 hp electric motor 50 hp electric motor

20 to 200 bfpd 20 to 200 bfpd 20 to 200 bfpd

Unconventional wells are drilled on a pad to minimize the footprint and reduce drilling cost. The pad
may contain 1 to 6 wells (Fig. 2). A well pad may also have a mix of flowing and pumping wells. A flowing
well may have a pumping unit installed before the well stops flowing (Fig. 3). This accelerates returning a
well to production once flow has stopped and artificial lift is required.

Figure 2—SRP Six well pad

SPE-181214-MS 5

Figure 3—SRP and flowing well

The economics of unconventional oil production drive the need to minimize capital and drilling cost.
This often results in less than optimal wellbore geometry for artificial lift when wells are drilled quickly.
Wellbore geometry is challenging for the SRP system because of deviations in the vertical section.
The SRP system must produce the well to the economic limit and wellbore geometry will have an impact
on future operating expense. Mitigation of tubing and sucker rod coupling wear will likely be the biggest
challenge during the life of the well. Estimated ultimate recovery (EUR) may also be impacted by wellbore
geometry.
Variation of wellbore geometry may be impacted by other factors such as surface location, off set
operators and reservoir structure. The lateral direction (Fig. 4) will follow the dip of the formation which
results in a toe-up or toe-down direction.

Figure 4—Unconventional wellbore geometry

A more complex wellbore may be drilled using a back build to maximize lateral length in situations
where the surface location is limited. Both toe up and toe down laterals create their own set of issues which
include multiphase flow, fluid dynamics and slugging.

Sucker Rod Pump System

SRP equipment is supplied by a variety of manufacturers. Selection of key system components Table 3 is
not considered a commodity purchase. To ensure optimal system performance each component and supplier
is evaluated and selected based on quality of service, application knowledge, technical support and cost.
6 SPE-181214-MS

Table 3—Key system components

Surface system Subsurface system

Pumping unit Bottomhole pump

Electric motor Sucker rods

Polished rod rotator Sucker rod guides

Critical service wellhead Tubing anchor

Variable speed drive Gas separator

This strategy provides Murphy the best technical solution utilizing business partners focused on reducing
the total cost of ownership. Vendors must be willing to collaborate rather than compete on behalf of the
company. The artificial lift team conducts quarterly service quality meetings with our strategic partners to
ensure alignment with the business and share lessons learned.

Variable Speed Drive

Murphy operates conventional and intelligent VSDs on all 550 SRP systems. The VSD facilitates real time
monitor and control and is connected to the desktop and enterprise monitor software via the SCADA system
(Fig. 5). The VSD operates in a number of fields with a generator (Fig. 6) and is easily connected to utility
power when available.

Figure 5—VSD connected to SCADA

Figure 6—125 kVA generator powering four VSDs

SPE-181214-MS 7

Onboard software turns the intelligent VSD into a computer which then manages pump speed, load,
torque, power and multiple constraint control. The VSD is configured to operate between a minimum and
maximum speed while maintaining a pump fill range or pump fill target.
A comparison between intelligent and conventional VSDs is outlined in Table 4.

Table 4—Comparison of intelligent and conventional

VSD-intelligent VSD-conventional

No load cell Load cell and cables required

No motor or crank sensors Magnetic motor and crank sensors required

Multiple constraint control Hard faults and limits

Power management No power management

Environmentally sealed cabinet External air cooling and filters

Simulator available No simulator

Single speed or automatic speed Single operating mode

Data displayed using operating units (SPM) Data displayed using non-operating units (Hz)

All faults configurable Faults either on or off

Remote access to software No remote access to software

Production Optimization
A VSD is essential to optimize production with the SRP in an unconventional oil well. The VSD manages
variability of inflow, gas interference and natural decline.
Traditionally, a fluid level datum point is collected, evaluated and then a decision is made to increase or
decrease the speed of the pump. A truck, technician, expensive equipment, training and good technique is
required to collect a single point.
This method of optimization has little relevance to producing a well in the Eagle Ford. Efforts have
been made to collect fluid level data and validate by comparing to dynamometer cards. The results were
inconsistent and not conclusive. Murphy's 550 operating SRP systems do not rely on fluid level data to
optimize production.
The dynamometer card (Fig. 7) is a plot of a suspected worn pump. Production data (Fig. 8) shows
a decrease of production during 15 days. The decision was made to increase the maximum speed of the
VSD with the intent to increase volumetric efficiency. The result of the speed change was an increase in
production of 30 to 40 bopd.

Figure 7—Dynamometer card worn pump

8 SPE-181214-MS

Figure 8—Oil rate and pump speed

Production optimization was completed from the desktop and did not involve an operator visit to the
wellsite.

Dynamometer Card
The dynamometer card is the primary diagnostic tool to monitor, optimize and troubleshoot the SRP
system. The dynamometer provides the plot of load and position within a single stroke. The VSD creates
a dynamometer card on every stroke. The dynamometer card is visible at the wellhead and remotely via
the enterprise monitor software.
Enterprise software (Fig. 9) collects and stores two dynamometer cards every 24 hrs. Dynamometer
cards can be polled on demand from the desktop when detailed analysis and troubleshooting is required.
Optimization technicians are not required to visit the wellsite to collect manual dynamometer cards.

Figure 9—Dynamometer card

Velocity Plot
Fig. 10 is a plot of sucker rod velocity, plunger velocity and pump speed of a normal operating pump.
Velocity data provides insight about system performance and can be used with the dynamometer card for
production optimization.
SPE-181214-MS 9

Figure 10—Velocity plot normal operation

Under normal pumping conditions, sucker rod velocity will look similar to a sinusoidal wave as speed
accelerates and decelerates at the top and bottom of the stroke. Plunger velocity will lag behind the sucker
rod velocity and has a significantly different speed profile. At point A and C the plunger velocity is 0 in./
sec during the transition from the up and down stroke. At point B the plunger has accelerated faster than
the sucker rods. On the down stroke, a similar event occurs.
Fig. 11 is velocity plot from a well which is pumping and flowing. This condition is commonly referred
to as "flumping". From point B to C, the plunger and sucker rod velocity are matching. At point A and C
plunger motion is still visible. This signature confirms the plunger is still attached to the sucker rod string.

Figure 11—Velocity plot pumping and flowing

System Design
The standard design followed "rules of thumb" and "industry best practice" which were developed mostly
in conventional applications. The lessons learned from the Eagle Ford have helped create a new set of
guidelines for SRP system design which are required to meet the demanding challenge of an unconventional
well.

Pump Landing Depth

The original strategy was to land the inlet of the pump 30 to 50 ft above the kick off point in the vertical
section. A deviation survey is used to identify the kick off point where inclination of the well bore increases
to build into the horizontal section. The dogleg severity (DLS) at the pump landing depth was not originally
considered. DLS may be defined as a rapid change in wellbore direction usually described as °/100 ft.
A deviation survey Table 5 indicates the kick off point of the lateral is located at 7,227 ft. The pump was
landed at 7,195 ft where the DLS is 3.74 °/100 ft. This pump operated for 504 days before failure. The top
10 SPE-181214-MS

section of an undamaged pump (Fig. 12) is as it normally appears after pulling out a well. The top assembly
of the pump with a cut valve rod guide and connectors was found (Fig. 13) because of side wear. During the
subsequent workover, the pump landing depth was raised to 7,164 ft. The new strategy is to land the pump
at a depth above the kick off point where the DLS is less than 2 °/100 ft.

Table 5—Wellbore deviation survey kick off point

Figure 12—Top pump assembly no wear

Figure 13—Top pump assembly cut from excessive side load

Solids
Solids are commonly associated with fracturing sand remaining from the reservoir stimulation. Most SRP
systems are producing trouble free with minimal impact from solids. A small group of wells have been
prone to sanding up when the reservoir releases solids into the wellbore. This may result in a seized plunger
and a wellbore packed with sand. Further evidence of solids can be observed during pump tear down. The
inside of the barrel and plunger show signs of scoring and wear. To date, no root cause failures of the pump
have been attributed to abrasive wear of the internal components. Therefore we do not use and special
configurations or modified pumps to handle sand.
Several wells have repeatedly sanded up. In these wells, a cleanout is required. Different cleanout methods
have been successfully used including a tubing bailer, coiled tubing and flushing with water and gel.
A strategy has been developed to minimize movement of solids into the wellbore. The reservoir builds
pressure after being shut in for a workover. To manage slugging at start up, pump speed is increased slowly
until the production is stable and no evidence of solids is observed. The VSD is then switched into auto
mode to begin to optimize production.
Other materials are found from samples collected during pump tear down and workover. Scale may form
on the sucker rods, tubing or pump components from water introduced to the well during the completion. The
practice of hot watering to remove paraffin from inside the tubing introduces the potential for contaminated
water which may contain bacteria. Various biocides are used to kill the bacteria. Treating bacteria is
challenging because of the variety and sources of the contamination. Unplanned chemical reaction may also
occur when reservoir fluid reacts with latent chemicals in the water.
SPE-181214-MS 11

Fluid Properties
CO2 and H2S are present in the Eagle Ford. CO2 concentrations vary between 0 and 10% across the three
fields. H2S concentrations vary from 0 to 7%. Field B has the highest concentration of H2S and CO2, Table 6.

Table 6—Fluid properties

Properties Field A Field B Field C

H2S 0 to 100 ppm 100 to 70,000 ppm 0 to 20 ppm

CO2 1 to 3 % 1 to 10% 1 to 3 %

Bottomhole temperature 285 to 310 250 to 275 225 to 250

API gravity 40 to 50 40 to 50 40 to 55

Gas specific gravity 0.75 0.75 0.75

Gas liquid ratio (scf/bbl.) 500 to 3,000 1,000 to 5,000 1000 to 20,000

Production Rate
The expected production rate from a well transitioning to artificial lift may be larger than the system capacity.
SRP systems were selected and designed based on a single artificial lift strategy for the life of the well.
A decision was made to not spend extra capital installing a larger SRP system to meet short term flush
production. The SRP system is designed to produce 20 to 200 bfpd which aligns to the decline curve during
the life of the well.

Sucker Rods
The sucker rod string design is based on expected loading while producing the well from the lowest possible
fluid level. The pump landing depth, tubing size, production rate, fluid properties, fluid level and pumping
speed all contribute to sucker rod load. An industry standard design software is used to model the SRP
system for a given application. The sucker rod string is designed for a maximum sucker rod load of 100
to 120% using a safety factor of 1.0. For wells producing in a corrosive environment the recommendation
is to use safety factor of .85 or .90.
A recommended standard sucker rod design and taper Table 7 for each field was developed.

Table 7—Sucker rod design and taper

Field A Field B Field C

.875 in. and .750 in. 1.0 in., .875 in., and .750 in. .875 in. and .750 in.

High strength High strength and KD D and KD

1.0 in. guided sinker rods 1.0 in. guided sinker rods .875in. guided sinker rods

Sucker Rod Guides

Sucker rod guides are installed to prevent contact with the tubing. Metal on metal contact of the sucker rod
coupling with the tubing wall eventually causes wear of the sucker rod coupling (Fig. 14) and a hole in the
tubing. Sucker rod guides have proved to be critical for SRP system reliability in highly deviated wells
12 SPE-181214-MS

Figure 14—Worn sucker rod coupling

The original strategy was to install guided sucker rods at locations in the wellbore where predicted side
load was greater than 40 pounds-force (lbf). The remaining sucker rods were installed with no guides which
followed another "rule of thumb" used in conventional SRP applications. As a result, many early life tubing
and sucker rod coupling failures were observed.
An investigation led to the creation of a new rod guiding strategy. The strategy incorporates a
recommendation from a proprietary sucker rod guiding software and data from previous failures noted in
the artificial lift failure data base. The sucker rod guide analysis considers wellbore geometry, side load and
diameter of sucker rod taper. A recommendation is made for the type of guide, placement on the sucker rod
and number of guides. Sucker rod configuration (Fig.15) options include 4, 6 and 8 guides/sucker rod. Most
sucker rods are guided with a minimum of 4/sucker rod. Placement of 6 or 8 guides/rod is recommended
in areas of high side load and DLS.

Figure 15—Sucker rod guide configuration

An alternate pattern may be used when sucker rods are installed with 8 guides/rod. This configuration
groups two sucker rod guides toward the sucker rod couplings. The remaining sucker rod guides are grouped
in pairs and evenly spaced along the sucker rod. This provides more protection closer to the sucker rod
coupling. The length of the guide, the orientation of the guide fins on the sucker rod and the fin style is
considered when selecting the rod guide.
Erodible wear volume (EWV) is the amount of sucker rod guide material available to wear before the
sucker rod coupling contacts the tubing. Larger diameter guided sucker rods have less EWV because of the
stand-off between the sucker rod coupling and tubing (Fig. 16). The EWV of 4 per guides on .750 in. sucker
rod is greater than 4 per guides on 1.0 in. sucker rod.
SPE-181214-MS 13

Figure 16—Sucker rod guide erodible wear volume

Sucker rod guides are designed to wear out in time. Selection of the correct model, material and
configuration is important. Manufacturing quality is also key to ensure a sucker rod guide can perform to
its technical limit. Careful inspection during work over operations identified potential quality issues with
certain sucker rod guides. As a result, guides became embrittled, cracked or missing from the sucker rod.
Guides were also found slide along the sucker rod and broke off in chunks. Several workovers occurred
because of premature failure of the sucker rod guide.
Fluid compatibility and temperature rating of the material may also be an underlying cause of failure.
Sucker rod guides may swell if left sitting in an environment with high temperature and high gravity oil.
Some wells have bottomhole temperature approaching the technical limit of the material. Common oilfield
chemicals such as paraffin and corrosion inhibitors must be considered when selecting a guide material.
Rod rotators are installed on all SRP systems. The theory behind rod rotators is rotating the sucker rods
will distribute wear evenly around the sucker rod guide to extend run life.
Another "rule of thumb" assumption is finding wear on one side of a rod guide indicates the rod rotator is
not working correctly. While this still may be true, the deeper Eagle Ford wells have developed distinctive
characteristics of guide wear which indicate another possible scenario.
During the workover, sucker rod guides are visually inspected as they are removed from the well. In the
upper section of the well, wear is distributed evenly around the sucker rod guide. In the middle section, rod
guide wear may be uneven or worn on one side. In the lowest section of the well, the guides may be severely
worn, broken, cracked, missing or mechanically damaged.
Rod rotators appear to be ineffective at rotating sucker rods in the lower section of a deep well. Friction
between the casing, tubing and sucker rods may limit rotation until there is sufficient buildup of torque
energy. Once the stored energy releases, mechanical damage may occur as the sucker rod guides violently
contact the tubing. This phenomena is commonly referred to as "stick slip" operation. The rod rotator has
been re-adjusted to increase the number of strokes required to complete one rotation of sucker rod string
and reduce the number of times "stick slip" operation occurs

Gas Interference
Gas interference was considered during the selection and design of the SRP system. Most conventional
vertical wells have the pump intake landed below the perforations which allows gas to separate in the casing
above the pump inlet. The casing acts as natural gas separator. In unconventional wells, the pump intake
is installed above the perforations. The current standard to decrease gas interference is to install a simple
conventional downhole gas separator. The gas separator (Fig. 17) is installed beneath the pump intake as
part of the tubing system.
14 SPE-181214-MS

Figure 17—Conventional gas separator

The gas separator is designed to decrease free gas from the fluid allowing liquid to enter the pump. Gas
interference is normal and expected when producing an unconventional well with an SRP system. Many
factors contribute to gas interference in an unconventional well which include:

• Pump intake installed above perforations

• Pump intake pressure below bubblepoint

• Turbulent flow in annulus

• Gas entrained fluid

• Oil gravity light

• Fluid is foamy

• Slug flow

• Gas liquid ratio (GLR) 500 to 20,000 scf/bbl

Six different models of downhole gas separators were installed.

• Long separator (40 ft) with diffused ports and long pre-installed dip tube

• Cyclonic model combination gas and solid separator with internal baffles

• Thin wall large diameter separator with pre-installed dip tube

• Packer model separator with complex internal and external flow path

• Helix model separator using centrifugal force for gas separation

• Traditional poor boy separator with perforated tubing sub and dip tube

Dynamometer cards were analyzed and gas interference was observed in all wells regardless of separator
model. The decision was made to install a low cost simple separator because no significant performance
differences were observed in any well.
SPE-181214-MS 15

Downhole Pump
Early installations used a variety of insert pump configurations. All pumps were bottom holddown cup type.
Common plunger sizes are 1.25, 1.50 and 1.75 in. The 1.75 in. pumps operate with a larger load on the
plunger and a shorter downhole stroke length compared to the 1.50 in. pump. This has resulted in operational
issues once the production rate declines and the fluid level is low. The effective pump stroke dropped to
40 in. and plunger load reach 10,000 lbs.
Different styles of pump were installed which included modified slippage, conventional and hollow valve
rod. Internal pump components used alternate pattern balls, Stellite lined cages, grooved plungers and cup
holddown. Variable slippage pumps were installed to address gas interference. In many situations this pump
was installed below 10,000 ft making pump space out difficult. This style of pump also underperformed
when the fluid level dropped and struggled to lift fluid to surface and is therefore no longer installed.
The current standard is a two stage hollow valve rod pump.

• 25-150-RHBM 24-6-2

• 25-150-RHBC-24-6-2

• 25-150-RHAC-24-6-2

Standard internal components are now designed to match fluid properties and production rate. Lessons
learned from the artificial lift data base are used to select the correct material and configuration of the pump.

Wellbore Geometry
Deviations in a well occur in three dimensions. Directional changes during drilling create a cork screw
path. The process of drilling an unconventional horizontal well includes the collection of real-time deviation
data to manage the trajectory of the well bore. The bottom hole drilling assembly is configured to make
directional corrections based on the deviation survey.
During drill pipe connection at surface, deviation data is collected at 100 ft to 200 ft intervals. The
deviation survey plots the distance a wellbore deviates from true vertical.

Lateral Section
A two dimensional plot of a lateral (Fig. 18) shows undulations which are common to a long horizontal
section. Lateral wellbore geometry is complex and may cause hold up of reservoir fluids and contribute
to slug flow.

Figure 18—Lateral deviation survey

16 SPE-181214-MS

Vertical Section
The looking north plot (Fig. 19) is a view of the vertical section of a well. This shows the deviation from
vertical in the east and west direction. This well deviates up to 30 ft to the west (point A) at several depths.

Figure 19—Looking north

The looking down plot (Fig. 20) is a view of a well from the surface location. The looking west plot
(Fig. 21) is a view of deviations from true vertical in a north and south direction. The well deviates 30 ft
north (point B).

Figure 20—Looking down

Figure 21—Looking west

Gyro Survey
The drilling survey does not always present an accurate representation of wellbore deviation. If unexplained
holes in the tubing occur or excessive sucker rod wear is found, a gyro survey is used to identify areas of
concern.
SPE-181214-MS 17

A plot (Fig. 22) compares the DLS from a deviation survey and a gyro survey. DLS data from the gyro
survey provides significantly more detail which was missed by the drilling survey. The gyro survey is more
practical and cost effective to collect high frequency deviation data compared to a drilling survey. Gyro
surveys can be run in casing or the tubing. No significant difference in DLS data was observed using either
method.

Figure 22—Dogleg severity comparison gyro survey vs drilling survey

Tubing Wear
Tubing wear mitigation has been a core area of focus for SRP pumped wells. Failure frequency of holes in
the tubing (Fig. 23) have been significantly reduced using conventional guided sucker rods and plain L-80
tubing. A strategy has been developed which includes tubing scans, gyro surveys and root cause failure
analysis.

Figure 23—Tubing of failure frequency

Wear happens when metal on metal contact occurs between the sucker rod coupling and the tubing which
results in a hole or split (Fig. 24). Additionally, high gravity oil with water has poor lubricity. This may
accelerate tubing wear therefore unguided sucker rods are not recommended for Eagle Ford SRP systems.
18 SPE-181214-MS

Figure 24—Split tubing-sucker rod coupling contact

During the workover a tubing scan is conducted to measure wall loss from sucker rod coupling wear
and corrosion pitting. The tubing scan establishes a base line of wear for the specific sucker rod design and
guide strategy installed in the well. A review of tubing scans has revealed a wear pattern which typically
occurs in a bell shape pattern. This confirms that replacing tubing one joint at a time or relying on a pressure
test has the potential to install badly worn tubing back in the hole.
The tubing replacement strategy includes:

• Replace tubing with wall loss greater than 30%

• Replace tubing above and below any laid down joints

• Install new tubing or yellow band in worn areas

The technology and equipment used for tubing scanning has a direct relationship to the data quality and
decisions about which tubing joints to lay down. Lower quality tubing scanning equipment only detects
wall loss around the circumference of the tubing which is mainly attributed to corrosion pitting.
The recommended scanning equipment detects both circumferential and longitudinal wall loss.
Longitudinal wall loss is attributed to sucker rod coupling wear on the inside of the tubing.
A review of the tubing system failures revealed holes in the tubing because of metal on metal contact
with the sucker rod coupling are distributed over all depths. This would be unexpected in a conventional
SRP well because most failures would occur near the bottom of the well. In the Eagle Ford the distribution
(Fig. 25) of tubing failures are split between 0 to 4,100 ft and 8,000 to 11,000 ft.

Figure 25—Distribution curve of location of tubing failure

SPE-181214-MS 19

The average run time of a tubing string which has failed is 346 days (Fig. 26). Tubing with short run
lives typically has an unguided sucker rod string or was rerun without a tubing scan which may result in
another unexpected failure.

Figure 26—Distribution curve of days to hole in tubing

A sucker rod without sucker rod guides will not always result in a hole in the tubing, but a correlation
can be made. To decrease the risk of a workover, guided sucker rod strings provide better reliability and
are more cost effective than lined tubing, hardened tubing or other types of sucker rod. The challenge with
installing modified tubing is to understand where to place in the wellbore.
Thermoplastic lined tubing was installed to attempt to mitigate tubing wear. This type of liner has
presented some significant operational challenges which include:

• Temperature limitation

• Cannot be inspected or scanned

• Material shreds and plugs up surface equipment

• Special handling procedures required

• Performance does not justify cost

Failure Data Base

Significant value was created to the business by conducting a detailed and focused effort on failure analysis.
Initially root cause of failure was not understood. This led to repeat failures as no corrective actions were
taken. As well count grew, failure frequency and total number of failures began to increase.
By June of 2014, a total of 150 systems were in operation. Failure frequency was 1.10 failures per well
per year. A data base was created and back filled with early failure information. Concurrently, new failures
were studied and added to the data base. A set of key performance indicators (KPIs) were developed based
on lessons learned.
As part of maintaining an accurate failure data base, monthly meetings are conducted at the pump shop
to review each tear down report. Worn, damaged and corroded parts are inspected and any samples of scale,
wax, solids and foreign material are collected. Samples are provided to the integrity team for further analysis.
20 SPE-181214-MS

Simple terms were developed for use in the tear down report to describe the condition of the pump. Like
new, slight wear, moderate wear and severe wear are used to describe the condition of the component. While
this detail may seem insignificant, it creates a common way to document the condition of the pump.

Key Performance Indicators

Key performance indicators have been created to measure run life improvement and reduction of failure
frequency. (Gardner et al. 2014) Reducing failure frequency has a direct impact to operating cost by delaying
the need for a workover. The goal is to design systems to operate for four years. Other data collected is used
to create a variety of scatter charts, bathtub curve, pie charts and distribution curves.

Subsurface Failure Frequency

Calculations were developed to measure subsurface system reliability. The first calculation counts the total
number of failures which have occurred during the previous 12 months divided by the total run days of all
wells during the same 12 months divided by 365.
A second calculation uses the total number of failures in the previous quarter and divides this number into
the cumulative operating days during the same quarter. The calculations measure historical change from the
previous 12 months as well as the current quarter. The subsurface failure frequency (Fig. 27) indicates a
reduction of failure frequency of 50% during the past 24 months.

Figure 27—Subsurface failure frequency

Subsurface Failure Frequency by Field

Plot (Fig. 28) shows early failure frequency by field was diverse. A wide range of operating conditions
including wellbore geometry and reservoir variability contributed to premature failures. Other factors such
as pump configurations also contributed to short run lives. The new standards have reduced failure frequency
by field which has normalize around more predictable values. Specific strategies have been develop for
each field.
SPE-181214-MS 21

Figure 28—Subsurface failure frequency

Bottomhole Pump
The scatter plot (Fig. 29) identified a reliability issue with the pump. A cluster of early life failures
were observed from December 2013 to December 2014. The days to failure was less than 100 days. An
investigation was undertaken to identify the root causes.

Figure 29—Bottomhole pump scatter plot

Several issues were identified during the study. A review of pump teardown reports identified integrity
issues with seat cups. The seat cups on the pump holddown were found to be torn, missing or hardened.
The torn and missing cups were determined to be from multiple seating and unseating events. The hardened
cups were associated with cups not being rated to the bottomhole temperature. Fluid wash and corrosion
was also observed on the pump holddown mandrel and as suspected, the corrosion extended to the carbon
steel pump seat nipple.
A decision was made to change from seat cups to a mechanical holddown with a stainless steel pump
seat nipple. This eliminates risk associated with seating and unseating during installation and issues with
bottomhole temperature and corrosion. A mechanical holddown relies of metal on metal to create a seal
rather than soft flexible material used in seat cups hence the improvement in reliability.
Another benefit of the mechanical holddown is the savings of operating cost when there is a need to
secure a well for an offset fracture stimulation. The process of securing the well involves unseating an
operating pump and placing a dry hole tree on the wellhead. After the stimulation is completed and pressure
in the well returns to normal, the pump can be reseated and the well returned to production immediately.
22 SPE-181214-MS

A pump with seat cups will need to be pulled out of the hole and replaced. The cost savings of mechanical
seat assemblies can be significant.
A few wells may warrant the use of a top holddown pump which use a seat cup assembly. Typically
these are shallow wells or wells which have a history of repeated sanding. To help improve the risk for top
hold down pumps, a high temperature seat cup is installed along with a stainless steel pump seat nipple
and stainless steel mandrel.

Sucker Rod
The sucker rod scatter plot (Fig. 30) identifies a group of failures with less than 300 days of run time.
Included in this group are approximately 17 wells with repeat failures. Small pits are difficult to see and
possibly missed by visual inspection. There is a possibility that sucker rods with small corrosion defects or
inclusions may inadvertently been run back in the well.

Figure 30—Sucker rod scatter plot

From the group of sucker rods which have failed in less than 300 days, corrosion-fatigue has been
identified as the most common root cause of failure. Investigation also found several failures were related
to handling. Sucker rod corrosion-fatigue problems appear to be isolated to certain wells located in clusters
across the field.
A batch treatment program has been initiated. The purpose of treating a well with chemicals is to provide
a coating to protect the steel when operating in a corrosive environment. During batch treating a mixture
of chemicals is pumped down the casing and the well is circulated up to 8 hours. The workover rig also
pumps corrosion chemical down the casing once all work is completed. Plans are to fine tune the chemical
program once specific causes have been identified and issues such as bacteria are quantified.
Chemical treatment programs are proven to protect the integrity of the casing, tubing, sucker rods and
pumps from failures related to corrosion. No chemical program can eliminate corrosion in every situation
but it can be part of a strategy to prolong run life and reduce total life cycle cost of the well as seen in
the Eagle Ford.

Tubing
The tubing scatter plot (Fig. 31) indicates the month installed vs. days to fail is improving. This is an
indication the sucker rod guide strategy and tubing scan program is improving system reliability.
SPE-181214-MS 23

Figure 31—Tubing scatter plot

Case Study 1 Tubing Wear

This case study is from a well which has experienced three tubing failures because of sucker rod coupling
contact. The MTBF is 251 days. An unguided sucker rod string was installed because predicted side load
over a majority of the well was less than 40 lbf (Fig. 32). DLS was less than 2°°/100 ft (Fig. 33). The
workover summary Table 8 outlines findings and attempts at corrective action at each intervention.

Figure 32—Side load

Figure 33—Dogleg severity

24 SPE-181214-MS

Table 8—Workover summary

Installation Pump depth Sucker rod guides Days to fail Cause of failure Fail depth Tubing scan
date

Production 11,186 ft No rod guides 188 Hole in tubing contact with 9,856 ft
period 1 coupling
New install
7-10-13

Production 11,223 ft Replaced slick rods .750 in. 99 Hole in tubing contact with 10,650 ft Yes
period 2 4/rod coupling Laid down 5 jts. with pitting
First repair 9,030 to 10,030 ft 1 split joint 9.856 ftno rod
2-11-2014 wear reported

Production 11,193 ft Replaced slick rods. 467 Hole in tubing contact with 10,720 ft No
period 3 875 in. 4 per/rod coupling Hydro tested tubing to 5000
Second 98 to 1,173 ft psi caused 3 jts. to split
repair Replaced and added.
6-19-2014 750 in. 4 per/rod
9,073 to 10,898 ft

Production 11,192 ft Replaced all rods operating Yes

period 4 Fully guided rods Laid down 107 joints with
Third repair 4 per/rod 5,875 ft rod coupling wear
11-25-2015 6 per/rod 1,500 ft
8 per/rod 3,700 ft

After 188 days of operation, the tubing failed because of sucker rod coupling contact. A tubing scan
identified pitting on five joints. One joint was found with split at 9,856 ft. The tubing scan report did not
identify any worn tubing. This did not correlate with visual wear reported on the sucker rods and couplings at
8,479 ft. A worn sucker rod was reported at 9,409 ft and worn sucker rod couplings observed from 9,409 to
11,188 ft. One thousand feet of guided .750 in. sucker rods (4 per/rod) were installed from 9,030 to 10,030 ft.
The tubing failed again after 99 days of operation. During the second workover, a split joint of tubing was
found at 10,650 ft. During this workover the tubing was not scanned and was hydro tested while running
back into the hole. Three joints split during the hydro test at 10,723 ft, 10,745 ft and 10,810 ft. Unguided
sucker rods were replaced with guided 4 per/rod from 98 ft to 1,173 ft. Guided .750 in. sucker rods with 4
per/rod were replaced and added from 9,073 ft to 10,898 ft.
During the third repair, the tubing was scanned using a better quality scanning process. The results of
the scan indicated significant sucker rod coupling wear in the tubing from 6,468 ft to 10,898 ft. A split
joint was found at 10,758 ft. Based on the results of the tubing scan, 107 joints of tubing from the lower
section of the well were laid down. A fully guided sucker rod string with 4, 6 and 8 guides per rod is now
installed in the well.
The lessons learned from this case study outline the importance of a good sucker rod guide strategy to
mitigate tubing wear. Predictive sucker rod modelling software does not always accurately represent the
side load and DLS in the wellbore. This is largely because of low frequency deviation survey data. The
tubing scan tool must be able to document corrosion pitting, sucker rod coupling wear, cracks, splits, holes
and defects. The tubing scan also provides a measure of sucker rod guiding effectiveness and a basis for
comparison during subsequent workovers.

Case Study 2 VSD Optimization

This case study outlines automatic production optimization by the VSD operating for 900 days. The VSD is
configured to operate between a maximum and minimum pump speed while maintaining a pump fill range.
Production optimization relies on the VSD software and eliminates need to collect fluid level data.
The plot (Fig. 34) shows the pumping unit operating speed and pump efficiency from the past 900 days.
Point A, maximum speed is set to single speed of 5.0 SPM. Point B, maximum pump speed increased to
SPE-181214-MS 25

6.5 SPM in single speed and pump fill drops dramatically. At Point C, the operating mode is switched from
single speed to automatic. The VSD now begins to adjust speed automatically to maintain the target pump
fill range. At point D, the VSD begins to adjust speed to match well inflow and slowly ramps speed down
over the 900 days.

Figure 34—VSD optimization 900 days operation

The plot (Fig. 35) displays the most recent 30 days of dynamometer cards. Reservoir inflow is stable
with periods of gas interference which are still being managed by the VSD after 900 days.

Figure 35—Dynacard most recent 30 days

Well Information

• 900 current operating days

• 8,800 ft pump depth, 1.50 in. pump

• 5.50 in. casing, 2.875 in. tubing

Case Study 3 Gas Interference

This case study demonstrates the ability of the VSD to manage gas interference. The strategy was developed
after downhole gas separators were found to be ineffective mitigating gas interference. The configuration
of the VSD was evaluated to understand if adjustments could be made to the speed profile beyond simple
control to maintain a pump fill range. A working knowledge of the type of VSD, programming and
configuration is required. "Out of the box" configuration and set up of a VSD are generally not intuitive
unless the goal is simple speed control.
The configuration of an intelligent VSD includes selecting a pump fill range from 85 to 92 %. The VSD
will adjust speed as pump fill moves outside the pump fill range. The minimum and maximum speed range
26 SPE-181214-MS

is set from 1.5 to 7 SPM. A pump fill measured above the set point range indicates additional fluid is present
at the pump inlet. The VSD gradually increases speed until the maximum speed is reached. When pump fill
drops below the pump fill range, pump speed is gradually decreased.
The goal is to have the speed of the VSD model a sinusoidal wave. This gives the reservoir time to
feed fluid into the wellbore and pump inlet. This allows the VSD to maintain steady state operation and
effectively manage gas interference while still trying to maximize production on each stroke.
Conventional VSDs require a different configuration to minimize abrupt speed change. The first
generation of conventional VSDs made aggressive speed changes when pump fill moved outside a fill target
of 85%. Typically these VSDs were found to be operating at the minimum speed of 2 SPM or at maximum
speed of 7 SPM. The drives were not configured to allow steady state operation. This resulted in production
instability, more gas interference and fluid slugging.
Problems have also arose during start up when a well began to flow. The VSD would calculate pump fill
of 100% which then resulted in a command to maximum speed. Conventional VSDs are now reconfigured
to decrease the rate of speed change and limit maximum speed to allow them to operate more efficiently.
In some cases, the pump fill target has been increased. This causes a delay in the speed change until the
pump inlet is almost completely full.
While studying dynamometer cards and gas interference, two distinct types of behavior were observed.
Type 1 dynamometer cards are defined by the VSDs ability to control speed within a pump fill range. These
wells generally have a moderate amount of gas in the fluid. A dynamometer card from a Type 1 well (Fig.
36) shows gas interference and the VSDs ability to modulate speed to allow inflow to quickly recover and
return to a full pump.

Figure 36—Type 1 Dynamometer card

Type 2 wells are defined by the VSDs inability to effectively control speed to maintain a pump fill range.
Gas interference may be observed on any stroke (Fig. 37). These wells generally have a high amount of gas
in the fluid. To reduce the effect of gas interference on Type 2 dynamometer cards, the VSD speed range
is limited to prevent overshooting. The VSD maximum speed is set to 3 to 4 SPM rather than 5 to 7 SPM
normally set on Type 1 wells. Limiting the speed range provides more time to recover from gas interference
and improve pump reliability.
SPE-181214-MS 27

Figure 37—Type 2 Dynamometer card

Case Study 4 Surveillance

Surveillance using enterprise software is conducted daily from the remote operations center (ROC) located
in Houston, TX. The ability to analyze multiple trends and over-lay dynamometer cards to measure system
performance is essential to maintaining production operations. The surveillance screens were configured
for all stakeholders to view data in the same format. This help to create situational awareness and manage
by exception.
The following series of plots outline the tools used to provide effective remote surveillance, production
optimization and troubleshooting of the VSD system. These tools eliminate the need for manual data
collection and minimize operator visits to the wellsite. Surveillance from the desktop helps identify the
source of production interruptions and direct field production optimization technicians to the areas of
concern.

Well Surveillance Screen

Plot (Fig. 38) provides a general overview of well and system performance and displays multiple trends on
a single plot. Changes become visible on trends which are updated every 15 minutes.

Figure 38—SRP surveillance screen

End Device Monitor

The end device plot (Fig. 39) is used for detailed analysis of pressure and temperature variation. Temperature
can be considered as a proxy for changes of flow. In many cases, temperature change can be correlated to
a pump failure, hole in tubing or parted sucker rod.
28 SPE-181214-MS

Figure 39—Surveillance pressure and temperature

Casing temperature data is used to identify a well which has begun to flow up the casing. Temperature
can also be correlated with the pressure to identify a subsurface failure such as a hole in tubing. Casing and
tubing pressure will equalize along with a change in the temperature profile when a hole in tubing occurs.

Pumping Unit Balance

The balance monitor plot (Fig. 40) is used to remotely check pumping unit balance and monitor changes
in pumping unit load. Parted sucker rods, out of balance and sticking sucker rods may be diagnosed. Any
change to load which results in a change of torque will be visible from the balance monitor. (Radzevicius
et al. 2012)

Figure 40—Surveillance pumping unit torque and balance plot

Dynamometer Card
This plot (Fig. 41) provides an overlay of multiple surface and downhole dynacards. This is the primary too
which measures sucker rod load, plunger load and pump fill. Dynamometer cards are collected twice daily on
all 550 SRP systems. The ROC reviews dynamometer cards daily and conducts a level one diagnostic which
looks for significant dynamometer card anomalies. Detailed analysis is completed by artificial engineers as
part of production optimization.
SPE-181214-MS 29

Figure 41—Surveillance dynamometer card

Recommended Practices
The goal is to take a long term view to improve run life, reduce failure frequency and mitigate repeat failures
which adds significant value to the business. These recommended practices are the result of significant
effort to identify the root cause of failure and then implement corrective actions.

Sucker Rod Pump

1. Land the pump above the kick off at DLS 2°/100 ft.
2. Ensure all components of the pump are correct material and specification
3. Install mechanical holddown where possible
4. For cup type holddown, ensure seat cups are rated to bottomhole temperature

Tubing
1. Scan tubing using a multi-directional tool
2. Replace tubing with wall loss greater than 30%
3. Run gyro survey if a repeated failure occurs

Sucker Rods
1. Install KD sucker rods for corrosive environments.
2. Install high strength rods as required. Ensure there is a corrosion treatment program
3. Inspect and replace worn sucker rod guides and couplings
4. Inspect and replace unguided sucker rods with coupling wear

Sucker Rod Guides

1. Install fully guided rod string using recommendation from software
2. Ensure number of guides, placement and material are adequate

Variable Speed Drive

1. Intelligent VSD required to manage inflow variability, decline and gas interference
2. Configuration of VSD to operate in gaseous fluid
3. Connect VSD to enterprise monitor software
30 SPE-181214-MS

Failure Analysis
1. Maintain data base and document root cause of failure
2. Review pump tear downs

Conclusion
SRP systems have proven to be economic and reliable to produce the Eagle Ford asset. During the past four
years 550 SRP systems have been installed. SRP systems are now fully integrated into Murphy's onshore
operations.
VSDs and enterprise monitoring software support the ability for remote operations, management by
exception and situational awareness. Production optimization is conducted daily with minimal operator
wellsite visits.
To ensure viability of SRP systems, it is important to understand the root cause of any failure. During
periods of rapid deployment, corrective action must be identified quickly and implemented. Murphy's
strategy is to continue to build on lessons learned to improve reliability, eliminate failures, reduce long term
cost and optimize production.

Acknowledgement
The authors would like to thank the following Murphy Oil personnel for their support. Eric Hambly,
Senior Vice President, US Onshore Operations, Paul Thayer, US Onshore Engineering Manager, Leslie
Armentrout, Reservoir Team Lead, Tom Krawietz, Well Works Team Lead, Molly Smith, Manager Onshore
Operations West, Francisco Garcia, Manager Onshore Operations East, Murphy's Artificial Lift Team, Well
Works Team and Reservoir Engineering Team and our vendor business partners. The authors would also
thank our joint venture partners for their contribution and support to the Eagle Ford asset.

References
1. Radzevicius, T., Clarke, F.; "Optimisation of Beam Pump Wells Utilising Variable Speed Drives
and Remote Monitoring Software", SPE Paper 148817, SPE Intelligent Energy International,
Utrecht, the Netherlands, 27-29 March 2012.
2. Gardner, Blair, Michael; "Reliability Improvement in Beam Pumps by Use of a Unique
Methodology Combining Real-Time Automation and Failure Analysis", SPE Paper 171372, SPE
Artificial Lift Conference and Exhibition North America, Houston, Texas, USA., 6-8 October,
2014.
3. Ron Martin, SPE, Jason Baihy, SPE Raj Malpani, SPE, Garrett Lindsay, SPE, W. Keith Atwood,
SPE, Schlumberger, "Understanding Production from Eagle Ford-Austin Chalk", SPE Paper
145117, SPE Annual Technical Conference and Exhibition, Denver, Colorado, USA, 30
October-1November 2011.