You are on page 1of 32


Table of Contents

INTRODUCTION ................................................................................ 1
CORROSION...................................................................................... 2
Electrochemical Corrosion .............................................................................. 2
Chemical Corrosion ......................................................................................... 3
Oxidation ......................................................................................................... 4
Galvanic Corrosion .......................................................................................... 4
Erosion-Corrosion ........................................................................................... 5
Hydrogen Embrittlement ................................................................................ 5
Corrosion Fatigue ............................................................................................ 5
FIELD SERVICE PROBLEMS ................................................................ 6
Conditions that Cause Sucker Rod and Coupling Failures............................... 6
Types of Sucker Rod and Coupling Failures .................................................... 6
DESIGNING TO REDUCE FAILURES................................................... 16
Mill Defects ................................................................................................... 16
Manufacturing Problems .............................................................................. 16
Handling Problems ........................................................................................ 17
Mechanical Damage ...................................................................................... 17
Improper Joint Makeup ................................................................................ 17
Bent Rods ...................................................................................................... 18
Guided Sucker Rods ...................................................................................... 19
Poor Pumping Conditions ............................................................................. 19
Corrosion Problems ....................................................................................... 20
Oxygen Corrosion .......................................................................................... 21
Problems Caused by a Stuck Pump ............................................................... 22
Hammering the Surface of Couplings ........................................................... 23
Thread Galling ............................................................................................... 23
Collection of Background Data and Selection ............................................... 25
Non -Destructive Testing (NDT) .................................................................... 25
Mechanical Testing ....................................................................................... 25
Selection and Preservation of Fracture Surfaces .......................................... 26
Macroscopic Examination of Fracture Surfaces ............................................ 26
Fracture Classifications ................................................................................. 27
REFERENCES ................................................................................... 29

1993-2016 Weatherford. All rights reserved. i



Historically, the first recorded incidence of corrosion was a Corrosion is a natural phenomenon that is not necessarily
problem encountered by British ships operating in the limited to metals. The effects of corrosion can be observed
Mediterranean early in the nineteenth century. Worms living every day, everywhere in the world; and the cost of its
in those waters would enter the wooden hulls and eat the damage to metal objects amounts to billions of dollars each
timbers until ships required dry-docking for replacement of year. Consider the following costs to an operator if a single
these wooden structures. The British decided to cover the rod string fails as a result of corrosive action:
ship hulls with thin sheets of copper and the worm problem
1. Downtime: lost production because the string can no
was solved, or so it appeared. Soon the copper sheets began
longer activate the pump
falling off the hulls, and the worm problem returned.
2. Workover: cost of a workover rig and the labor to pull
The steel nails holding the copper had disintegrated where
the tubing and rod string
the copper and steel were in contact, and no one could
explain the reason. We now know that a galvanic action 3. Replacement: cost of a new string of rods and possibly
occurs between dissimilar metals in sea water. Using copper the cost of replacing tubing and the downhole pump
nails resolved the problem.
4. Work: labor of company personnel who could be doing
Last century, scientists began to recognize the tremendous other things
scope of corrosion and the cost associated with failed
Multiply those costs by the number of corrosion failures in
equipment. Of note was the realization that stray current
rod strings in one year and the total cost to operators
from street car railways was damaging and even destroying
becomes enormous.
underground metal structures and communication cables.
Today the mechanics of corrosion are understood and its
behavior can be controlled. In some cases it can be
completely eliminated by following proper procedures. This
guide focuses on the types of corrosion related to sucker
rods, the causes, and ways to reduce sucker rod failures.

1993-2016 Weatherford. All rights reserved. 1



Corrosion can be defined as the deterioration of a Electrochemical Corrosion

substance or its properties because of a reaction with its
environment. The substance we will consider is steel. Steel Most corrosion in the oil field is attributable to the presence
is an iron compound with alloying elements such as carbon of water in large or small amounts. Corrosion in the
and manganese. presence of water is an electrochemical process. Please keep
this in mind as we discuss other types of corrosion.
The driving force that makes metals corrode is a natural
phenomenon. Iron, from which steel is produced, is found in Figure 1 shows that steel is not homogeneous. It is a mixture
combination with other elements, and these are called ores. of iron, carbon, and other alloying elements; but for
Any iron-based product in a usable form is in only a purposes of illustration, lets consider only the iron and
temporary condition and eventually, if let unhampered, will carbon elements. Some of the carbon is dissolved in the iron,
return to its original form (an ore). and the balance exists as iron carbide.

Left unprotected in an atmosphere, a metal will release

energy as it combines with other elements and returns to
its natural state as an ore. The release of energy and its
attendant combination with other elements to form ore
is corrosion.
One of the most obvious examples is a steel object left
exposed to weather. It will begin to rust and, if left
undisturbed, will completely deteriorate. Rust is a
combination of iron and oxygen, in which the iron gives up
its energy and returns to its natural state.
There are multiple types of corrosion:
Electrochemical corrosion
Chemical corrosion
Sweet Figure 1: Photomicrograph showing a pearlite ferrite
microstructure. Pearlite is a mixture of cementite (Fe3C); Ferrite is
Oxidation almost pure iron.
Galvanic Iron carbide (Fe3C) has a lower tendency to corrode than
Corrosion-erosion pure iron (Fe), thus with an electrode and in the presence of
an electrolyte (such as water or salt water), the two different
Hydrogen embrittlement compositions will complete an electrical circuit and current
Corrosion fatigue will flow (Figure 2).

In most instances more than one type of corrosion

contributes to a failure.

1993-2016 Weatherford. All rights reserved. 2


The area corroded is always the anode. The cathode

remains unaffected.
The affinity of iron in solution for oxygen is greater than the
affinity of hydrogen for oxygen. Hydrogen gives up one
electron when it goes into its ionic state. Oxygen gains two
electrons and becomes negatively charged.
In a simplified form, the reaction is as follows:
Fe + H2O FeO + H2

Figure 2: Schematic showing current flow between iron and iron Iron Atom Water Iron Oxide Hydrogen Gas
carbide grains with resulting corrosion of the iron

Some definitions are necessary to describe the Fe+ + H+H+O FeO + H2

electrical circuit: Positive Iron Negative Oxygen Iron Oxide Hydrogen Gas
Positive Hydrogen
1. Anode. The anode is that portion of the metal surface
that is corroded. The positively charged iron ion and the negatively charged
oxygen ion combine to form iron oxide. The two positively
2. Cathode. The cathode is the opposite side of the cell
charged hydrogen ions then form a molecule and escape at
that is unaffected by the current flow and therefore does
the cathode as a gas.
not corrode.
3. Electrolyte. The fluid that transmits the current between Chemical Corrosion
anode and cathode. Sour
When iron corrodes, the iron dissolves, goes into solution, Hydrogen sulfide is very soluble in water. When both
and gives up two electrons. Since an atom of iron contains hydrogen sulfide and water are present in well fluids, sulfuric
an equal number of protons (positively charged particles) acid (H2SO4,) is formed. Although other reactions may be
and electrons (negatively charged particles), the iron in involved, simply stated, the final reaction is the formation of
solution is called an ion. The iron ion is positively charged. iron sulfide (FeS).
Fe Fe2+ + 2e H2S + Fe FeS + H2
Iron Atom Iron Atom Electrons Hydrogen Sulfide Iron Atom Iron Sulfide Hydrogen Gas

Although the cathode remains unaffected, another chemical Iron sulfide is a black scale that clings to the surface of the
reaction takes place at the cathode. The electrons left metal and is cathodic to the iron. It accelerates the corrosion
behind at the anode when iron went into solution travel to in the locality of the scale and causes deep pitting.
the cathode through the solid metal. Hydrogen in water has
an affinity for those electrons and consumes them.
The term sweet corrosion refers to corrosion caused by the
2H+ + +2e H2 presence of carbon dioxide (CO2) in a producing well. Carbon
Hydrogen Ions Electrons Hydrogen Gas
dioxide easily dissolves in water to form carbonic acid.
CO2 + H2O H2CO3
Carbon Dioxide Water Carbonic Acid

An electrochemical reaction occurs and the iron replaces

hydrogen to form iron carbonate.
Fe + H2CO3 Fe2CO3 + H2
Iron Carbonic Iron Carbonate Gas

The end result is pitting. The severity of the pitting is

determined by many factors, such as pressure, temperature,
and the amount of CO2 present.
Figure 3: Schematic showing basic current and electron flow in
localized corrosion cell at the surface of ferrous metals immersed in Both sour and sweet corrosion can occur in the same well.
an electrolyte

1993-2016 Weatherford. All rights reserved. 3


Oxidation When the three conditions listed above are met, the anodic
(most active) metal will deteriorate while the cathodic (less
Corrosion of metal by oxidation is the ordinary rust observed active) metal will be relatively unaffected. Figure 4 indicates
on any unprotected piece of steel. It can only occur in the the relative anodic and cathodic properties of some
presence of water, and then only if dissolved oxygen is commercial metals and alloys in sea water. From their
present. However, it will also occur in sea water and salt positions on the scale in Figure 4, it becomes apparent that
solutions. We will limit our discussion to salt water, a fluid steel nails will be destroyed by the more noble copper. The
that is prevalent in pumping wells. scale also explains why gold is usually not found in chemical
Water solutions rapidly dissolve oxygen from the air, and combination with other elements.
this is the source for the required oxygen in the corrosion Active Magnesium
process. Normally oxygen is not found in well fluids; but if it or Magnesium Alloys
is introduced by any method, rusting will continue as long as Anodic Zinc
oxygen is present. The simplified chemical reaction is: Galvanized Steel
Aluminum 1100
H2O Aluminum 2024 (4.5 Cu, 1.5 Mg, 0.6 Mn)
4Fe + 3O2 4Fe2O3 Mild Steel
Iron Oxygen Iron Oxide Wrought Iron
Cast Iron
For an example of the effect of oxygen on steel, consider the 13% Chromium Stainless Steel
photos taken of the ocean liner Titanic that lies at the Type 410 (active)
bottom of the Atlantic Ocean. Why has the hull not corroded 18-8 Stainless Steel
away? Because at those depths, no oxygen exists. TypeTin Solders
Galvanic Corrosion
Galvanic corrosion is very rare in the oil field. Three Muntz Metal
conditions are necessary for galvanic corrosion to occur: Manganese Bronze
Naval Brass
Electrochemically dissimilar metals must be present.
Nickel (active)
The metals must be in electrical contact. 76 Ni, 16 Cr, 7 Fe Alloy (active)
60 Ni, 30 Mo, 6 Fe, 1 Mn
The metals must be exposed to an electrolyte. Yellow Brass
Electrochemical dissimilarity refers to the amount of Admiralty Brass
energy stored by the metal when it is removed from its Red Brass
natural condition. Copper
Silicon Brass
An example of metals in electrical contact is the steel nails
70:30 Cupro Nickel
that held copper sheeting on the hulls of British ships in salt
water. A potential can be created between any two
Silver Solder
dissimilar metals and even between different types of steels.
Nickel (passive)
The third condition for galvanic corrosion relates to the 76 Ni, 16 Cr, 7 Fe
strength and type of electrolyte and the presence or absence Alloy (passive)
of oxygen. 13% Chromium Stainless Steel
Type 410 (passive)
Figure 4: Galvanic series of some commercial metals and alloys in
sea water

1993-2016 Weatherford. All rights reserved. 4


Erosion-Corrosion Corrosion Fatigue

This type of corrosion is caused by corrosive fluid, impinging Thus far we have primarily considered the mechanics of
flow, or turbulence of the fluid upon metal surfaces. corrosion alone. However, corrosion and fatigue, acting in
Impingement can be caused by solid material or gas bubbles concert, are the cause of many sucker rod failures.
entrapped in the fluid being pumped. These materials,
Cyclic stressing of sucker rods does not typically cause a
moving against a metal surface, tend to abrade or erode the
failure of the sucker rod string if the stresses are limited to a
steel. The erosion takes the form of elongated pitting or a
level below the endurance limit of the steel. Environmental
deep groove. The rapid movement of corrosive fluid
conditions are extremely important in corrosion fatigue. Salt
removes a protective scale and exposes the underlying
water is a corrosive medium. When dissolved gases such as
metal, which accelerates corrosion.
hydrogen sulfide (H2S), carbon dioxide (C02), and free oxygen
Turbulence is a major factor in this type of attack. When a are present in salt water, corrosivity increases and fatigue
liquid flows over a metal, there is usually a critical velocity life decreases.
below which impingement does not occur, but above which
Usually fatigue begins at the rod surface as a pit in the
impingement rapidly increases. To illustrate this
metal, a nonmetallic inclusion, or some other steel
phenomenon, consider a curve in a lazy stream. As long as
defect. However, it may begin with some kind of
the stream contains little water and the flow is gentle, the
mechanical damage.
banks remain unchanged. Increase the water volume and
the stream becomes more turbulent, thereby eroding Though the stresses may be uniform over the balance of the
the bank. rod, the stresses induced at the bottom of a pit will be
considerably higher; and as corrosion continues to attack,
The break out of gases in low-pressure zones can
the cross section is reduced. As stresses across that area
impinge on the surface of a sucker rod and remove any
increase, a fatigue failure results.
protective coating.
This will be discussed later in this guide.
Hydrogen Embrittlement
We have discussed chemical and electrochemical reactions
in which positively charged hydrogen ions in an electrolyte
are expelled from the system as hydrogen gas. However, all
hydrogen may not leave. It is possible, even most probable
in a sour well environment, that some of the hydrogen will
enter the steel as either atomic or molecular hydrogen and
diffuse into its structure. Once absorbed, the hydrogen
molecules build pressure at crystallographic vacancies or
discontinuities, such as voids, which generate microscopic
cracks. This results in a brittle failure of the steel at stress
levels considerably below the yield strength of the steel.
The more acidic the well fluid, the more susceptible the steel
is to hydrogen embrittlement. Since hydrogen
embrittlement is related to the chemical and
electrochemical reactions we described in sour wells,
the term sulfide stress cracking is also used to describe
this phenomenon.

1993-2016 Weatherford. All rights reserved. 5


Field Service Problems

In this section we will discuss field problems that you will

Handling problems
encounter with sucker rods and sucker rod couplings. We
will also show examples to illustrate each mode of failure. Improper makeup procedures
Bent rods
Failure prevention of any component starts with a
Mechanical damage
knowledge of failure mechanisms. To prevent failures, it is
necessary to understand the many ways a product can fail. Inhibitors
Each failure can be considered a valuable specimen from
Lack of inhibition
which to extract as much information as possible. The
Poor inhibition program
information can be used to improve product quality and
skills in the application of sucker rods. Understanding Manufacturing defects
product failures and being able to extend field service life is
Mill defects
important to you, the salesman or the field engineer, and to
Forging defects
the user who will benefit by receiving longer service from
Threading defects
the equipment and by paying less for lifting.

Conditions that Cause Sucker Rod and Coupling Types of Sucker Rod and Coupling Failures
Failures Metallurgical investigations of failed sucker rods typically
reveal a fatigue pattern. The fatigue crack originates at the
Examination of sucker rod failures reveals several enemies to point of highest stress, and the stress peaks occur at the
dependable rod string operation. S.M. Bucaran, et aI, surface. Therefore, the fatigue failures are surface failures
presented the paper (No. 55) NACE/72 "Proper Selection, and the damage, or intrusion, is a sharp notch or roughness
Handling, and Protection of Downhole MaterialsA Practical and a stress-raiser that starts the crack.
and Economical Approach" in which they stated that rod and
coupling failures can be classified simply as caused by wear, Breaks in sucker rod pins have a similar appearance to
corrosion, mechanical causes, or mishandling; and plenty of fatigue failure, but they usually occur if the joint is not made
cases involve two or more of the basic causes .
1 up properly, a condition referred to as loss of displacement.
A loss of displacement occurs when a connection loosens
Here is another list of causes for sucker rod and downhole. This can be caused by improper makeup
coupling failures: procedures, applications issues, or mechanical damage.
Overload condition Coupling breaks will occur as a result of wear, corrosion, or
Application issues loss of displacement. Hydrogen sulfide embrittlement
failures have occurred in couplings with hardness greater
Rods in compression than 23 Rockwell C.
Rod wear
Pumping speed We will define and discuss four modes of sucker rod and
Stuck pump coupling failures:
Incomplete pump fillage Fatigue: the effects of stress concentration and corrosion
Corrosion problems Corrosion: pitting versus uniform attack, and the
Water floods damaging effects of hydrogen sulfide embrittlement
Sour gas Wear: its effects on service life
CO2 injection
Tensile failure: not a service failure but an overload
condition caused by carelessness

1993-2016 Weatherford. All rights reserved. 6


Fatigue The process of fatigue consists of three stages (Figure 6):

The word fatigue may suggest that metals become tired 1. Initial fatigue damage leading to crack initiation
from supporting a load for a long time. Certainly metals do
not tire in a biological sense; nor do they deteriorate as a 2. Crack propagation until the remaining uncracked
direct result of supporting a constant load. However, they do cross-section of a part becomes too weak to carry the
fracture in a brittle manner when subjected to cyclical imposed loads
loading that varies sufficiently in intensity, even though one 3. Final, sudden fracture of the remaining cross-section
such cycle produces no detectable effect.
Fatigue is the progressive, localized, permanent structural
change that occurs in a material subjected to repeated or
fluctuating stresses that have a maximum value less than the
tensile strength of the material. Fatigue fractures are caused
by the simultaneous action of cyclic stress, tensile stress, and
plastic strain, all three of which must be present.
Although these three conditions are sufficient to cause
fatigue, a host of other variablessuch as temperature,
crystal system of the metal, grain size, environment
(corrosive or otherwise), metallurgical structure, and stress
systemalter the conditions for fatigue. Figure 5 illustrates
the three general types of fluctuating stresses that can
Figure 6: Relative lengths of the various stages of a sucker rod
cause fatigue. body failure

Consequently, fatigue cracks are classified as brittle cracks,

as identified by little or no evidence of ductility in the area
adjacent to the second stage cracking. There is no necking or
shear lips on the fracture as shown in the final stage of
failure. Fatigue failures are always normal (90) to the
applied stress. Chevron marks, or a herringbone pattern,
point to the fracture origin.
Fatigue beach marks, clamshell marks, or families of
striations may be observed with the unaided eye
(Figure 7). The familiar beach marks can be observed
macroscopically with a light microscope or under the
electron microscope (Figure 8).

Figure 5: Typical fatigue stress cycles

Case 1 represents an idealized situation wherein stress

fluctuates in a sinusoidal fashion from tensile to
compression and the net resultant stress is zero. This is the
most common form used to study fatigue in a laboratory,
but it is also approached in service by a rotating shaft
operating at a constant speed without overload.
Figure 7: The three stages of fatigue of a sucker rod body failure
In Case 2, we have the sinusoidal stress form but the
resultant, or mean, stress is not zero.
Case 3 shows an irregular or random stress cycle.

1993-2016 Weatherford. All rights reserved. 7


concentrated loads at a surface discontinuity. Such

concentrated stresses produce local plastic strain (stress
exceeds yield strength) that is critical in cyclic loading.
In practice, prediction of the fatigue life of a material is
complicated because the fatigue life is very sensitive to small
changes in loading conditions, local stresses, service
environment, and local characteristics of the material.
However, the endurance limit of any material under cyclic
stress is lower than its strength under static load. As load
decreases, the number of cycles to failure increases until at
some load, the number of cycles to failure becomes so large
that we need not fear failure. Normally the endurance limit
for sucker rods is defined as the maximum stress that the
Figure 8: Field failure with visible beach marks string will withstand without failure after ten million or more
cycles of stress.
Fatigue beach marks are clearly evident with little or no
magnification for lower strength, more ductile materials and Generally, when loads are low, only one crack is generated.
for a variety of strength ranges under conditions of low load, Conversely, multiple cracking is a sign of high loads. In addition,
high-cycle stresses. The distinguishable concoidal markings the ratio of prior crack area to total cross-sectional area gives
usually represent points of variation in the load information about the magnitude of stress at final rupture. With
environment. For higher strength materials and for high-load the use of an electron microscope, direct measurements of
fatigue, the use of an optical or electron microscope is often striation spacing offers good insight into the stress environment
required. With the close examination of a fatigue fracture during crack growth.
face, much or all of the following information regarding the
The preparation of a metallographic specimen containing a
crack can be determined:
crack is often helpful in identifying the crack mechanism.
1. Point(s) of crack nucleation Fatigue cracks are invariably transgranular or transcrystalline
and sometimes are branched. Fatigue striations are not
2. Direction of crack growth
evident, however, on profile.
3. Size of prior crack
Bending fatigue failures can be divided into three
4. Relative magnitude of stress classifications: one-way, two-way, and rotary. The fatigue
crack formations associated with the type of bending load
5. Direction of loading (axial, bending, reverse bending, etc.)
are shown in Figure 9.
Fatigue failures occur at apparently safe values of
predetermined, calculated average stress over the cross-
section of a sucker rod or coupling, and failure is caused by

Stress Condition
Case No Stress Concentration Mild Stress Concentration High Stress Concentration
Low Overstress A High Overstress B Low Overstress C High Overstress D Low Overstress E High Overstress F

One-way bending load

Two-way bending load

Reversed-bending load-
rotation load

Figure 9: Fracture appearances of fatigue failures in bending by Dr. Charles Upson, "Why Machine Parts Fail," Penton, Cleveland 13, Ohio

1993-2016 Weatherford. All rights reserved. 8


API Bulletin RP11BR "Recommended Practice for Care and 140

Handling of Sucker Rods" recommends using a modified 130 cases

Goodman diagram (Figure 10) for determining the allowable

range of stress for a string of sucker rods in noncorrosive service. 120
50% ratio
Minimum Tensile Strength T 110

Endurance limit, 1,000 psi, Sn



2 SI 1.75 70

l e

llow Normal for

T =A

polished specimens

3 SA




Severely notched specimens


45 30
0 0
SA = (4T + M S ) SF
1 20
SA = (0.25T + 0.5625 S ) SF 2
Corroding specimens

SA = SA SMIN 3 10
When: 40 60 80 100 120 140 160 180 200 220 240 260
T SA = Maximum available stress, PSI
3 SA = Maximum allowable range of stress, PSI
Tensile strength, 1,000 psi, Su
M = Slope of SA curve = 0.5625
SF =
Minimum stress, PSI (calculated or measured)
Service factor
Figure 11: Generalized relation of ultimate tensile strength (Su) and
T = Minimum tensile strength, PSI fatigue strength (Sn)

Figure 10: Modified Goodman diagram for allowable stress and Figure 12: Stress concentration factors for various surface
range of stress for sucker rods in noncorrosive service conditions. This graph is a derating factor for influence of surface
conditions on fatigue.
Figure 11 shows a number of variables that influence the
fatigue properties. Note that surface notches and corrosion Sn = 0.5(Su)(C1)(C0)(CS)
reduce fatigue strength. Figure 12 shows stress Sn = Endurance limit
concentration factors for the various surface conditions.
Su = Ultimate strength
Hardness, Bhn
Sn/Su = 0.5
120 160 200 240 280 320 360 400 440 480 520
1.0 C1 = 1.0, 0.9, or 0.58 depending on whether the load is
by bending, axial, or torsion, respectively
Fine-ground or
commercially polished
C0 = A size factor, usually taken at 1.0 for diameters
less than 0.4 inches and at 0.9 for diameters
Surface factor CS

0.7 Machin
ed between 0.4 and 3.0 inches
0.6 CS = Surface factor, which varies for the surface
conditions shown in Figure 12

0.4 Ho
t- r olle
0.3 As f
o r ged
0.2 Corroded in
tap water
Corroded in salt water
60 80 100 120 140 160 180 200 220 240 260
Tensile strength Su, ksi

1993-2016 Weatherford. All rights reserved. 9


Figure 13 shows the importance of limiting the system to Sucker rods have a hot-rolled surface (note derating factor
low and intermediate hardness and indicates the importance for hot-rolled materials in Figure 12). For a sucker rod with
of residual stress in fatigue; that is, materials with a greater 100,000-psi tensile strength, the endurance limit for
endurance limit are higher in carbon and thus have a complete stress reversals would be approximately 50,000 psi
higher temperability. for polished test specimens and 30,000 psi for test
160 specimens with a hot-rolled surface. A salt water
H-11 environment further reduces fatigue strength.
The API modified Goodman diagram (Figure 10) has been
140 adjusted to indicate a derating for a hot-rolled, shot-blasted
H-11 surface and a substantial safety factor of 2. The adjustment
130 or derating for corrosion (service factor) is selected by each
A user as his experience indicates. Figure 14 indicates typical
service factors for various environments.
Endurance Limit1,000 psi

110 B

90 E, F


70 A = SAE 4063
B = SAE 5150
C = SAE 4052
D = SAE 4140
E = SAE 4340
50 F = SAE 2340

20 30 40 50 60

Rockwell C Hardness
Figure 13: Relation of hardness and fatigue strength for
several steels

Corrosiveness/Derated Service Factor

Environment None Mild Moderate Severe
1.0 0.9 0.8 0.7
H2S 0% X < 10 ppm (0.001%) 10 ppm < X < 100 ppm (0.001% - 0.01%) 100 ppm < X (0.01% and greater)
CO2 0% X < 250 ppm (0.025%) 250 ppm < 1500 ppm (0.025% - 0.15%) 1500 ppm < X (0.15% and greater)

Figure 14: Table of derating factors (NACE Standard MR0176-2000)

1993-2016 Weatherford. All rights reserved. 10


The API modified Goodman diagram cuts off at neutral (zero) Exercise #1
rather than at reversed stress because the rod string is
Assume a string of API Grade C sucker rods with a minimum
susceptible to column buckling and should not be put
tensile strength of 90,000 psi (T) is being used at a minimum
into compression.
downstroke stress of 10,000 psi (Smin). At what peak polished
To avoid exceeding the yield strength (plastic range), the rod stress (SA) can we operate this string in noncorrosive
upper boundary of the shaded portion in Figure 10 service (SF = 1)?
represents approximately 58% of the ultimate tensile
strength, which is less than the yield strength. = +
The API version of the Goodman diagram helps in visualizing SA = (90,000/4 + 10,000 x 0.5625) 1
the failure strength of a sucker rod string with a variety of
loads. The relationship is shown in the shaded area of SA = (22,500 + 5625) 1
Figure 10: The top line on the shaded area represents the SA = 28,125 PSI in noncorrosive service
maximum stress and the bottom line represents the
Converting to load for different size top rods:
minimum stress with which the sucker rod can be loaded
without fatigue failure in a noncorrosive environment. As L = Load
the stress range (maximum stress minus the minimum
SA = Maximum available stress, psi
stress) is reduced, the maximum stress and minimum stress
can be increased. When the stress range is reduced to zero, Arod = Cross-sectional area of sucker rod body
the load is static. To operate in the safe range, it is necessary
L = SA (Arod)
to determine the maximum allowable loading on the basis of
calculated or measured minimum allowable loading. 5/8 in. 28,125 psi .307 in2 = 8,634 lb
Example 3/4 in. 28,125 psi .442 in2 = 12,431 lb
(Refer to notes covering calculations and Goodman diagram 7/8 in. 28,125 psi .601 in2= 16,903 lb
in Figure 10).
1 in. 28,125 psi .785 in2 = 22,078 Ib
Example Allowable Sucker Rod Stress Determination Using
Exercise #2
Range of Stress
What is the peak polished rod stress at which we can
Calculating maximum allowable stress:
operate in H2S environments (remember SF for H2S is 0.80)?

= +
4 = +
= + 0.5625 = (90,000/4 + 10,000 0.5625) 0.80
Where: = (22,500 + 5625) 0.80
SA = Maximum available stress, psi SA = 22,500 lb stress in H2S service
M = Slope of SA curve = 0.5625 Exercise # 3
T = Minimum tensile strength, psi For API Grade C, given a peak polished rod stress of 35,000
Per API 11B T = : psi, calculate minimum allowable stress (Smin).

Grade K = 90,000 psi

= 0.5625
Grade C = 90,000 psi
= (35,000 0.25 90,000) 0.5625
Grade D = 115,000 psi
= (35,000 22,500) 0.5625
Smin = Minimum stress, psi (calculated or measured)
= 22,222 psi in noncorrosive service
SF = Service factor (noncorrosive = 1, H2S = 0.80)
Calculating minimum allowable stress Smin:

= 0.5625

1993-2016 Weatherford. All rights reserved. 11


To convert for different size top rods in corrosive service: cracking, but small pits with sharp roots are logical locations
2 for crack initiation.
5/8 in. 28,125 psi 0.80 0.307 in. = 6,908 lb
There are several types of downhole corrosion and they
3/4 in. 28,125 psi 0.80 0.442 in.2 = 9,945 lb
affect sucker rods differently. However, all pumping wells
7/8 in. 28,125 psi 0.80 0.601 in.2 = 13,523 Ib produce fluids that are corrosive to some degree. Figures 15
and 16 show fatigue cracks in corrosion pits, one from a sour
1 in. 28,125 psi 0.80 0.785 in.2 = 17,663 lb
well (H2S) and the other from a sweet well (CO2).
We have defined and discussed metal fatigue. We know that
metals under cyclic loading have a limited useful strength in
a noncorrosive environment and that the endurance limit
(fatigue strength) is primarily dependent on tensile strength.
Sucker rods that operate below the endurance limit in a
noncorrosive environment will withstand an infinite number
of pump strokes (stress reversals) before failure. These
factors affect sucker rod fatigue:
The range between minimum and maximum tensile
Figure 15: Start of fatigue cracks in corrosion pits in a sour
stress: the wider the range, the lower the number of well (H2S)
cycles to produce a fatigue failure
Pitting, stress cracking, and severe uniform corrosion
The environment in which the sucker rod is used
The environment has a very pronounced effect on fatigue
strength. In laboratory tests, samples tested in air and in tap
and salt water showed progressively lower fatigue strength.
The more corrosive the environment, the lower the fatigue
strength becomes. Refer to Figure 11 for derating factors for
various surface conditions and environments.
Corrosion combined with cyclic stress is more damaging than
either corrosion or fatigue alone. The part played by
corrosion in this type of degradation is extremely important.
Without corrosion, fatigue failures would be greatly
reduced, or very likely there would be no fatigue cracking
except in the few cases in which fatigue failures were caused
by mechanical notches. However, the presence of corrosion
lowers the endurance limit of steel. There is actually no real Figure 16: Start of fatigue cracks in corrosion pits of a well
containing CO2 (sweet well)
endurance limit in a corrosive environment; whether the
endurance will be higher or lower is dependent on the salt
content of the water and the presence of oxygen, carbon
dioxide, and hydrogen sulfide gases.
Corrosion-fatigue failures can be considered similar to notch
fatigue failures. The difference is that the corrosion-fatigue
crack starts at the root of the corrosion pit, but notch fatigue
starts at the root of mechanical defect.
Corrosion-fatigue can be considered as a type of notch
fatigue in which a point of stress concentration has been
formed by a corrosion pit, and the fatigue progress is
accelerated by the action of corrosion. Not all corrosion pits
produce cracks at the root. General-type pitting typically
does not produce sharp notches, but rather flat or rounded-
bottom pits. This type of pitting is less likely to produce

1993-2016 Weatherford. All rights reserved. 12


Development of a Modified Goodman Diagram for +

Sucker Rods
A fatigue test consists of placing a specimen in the fatigue Smax
machine, subjecting the specimen to a defined load (typically

reverse bending by rotation), and running the machine until

the specimen fails or until it has run in excess of ten Smin

million revolutions.
The process for developing a modified Goodman diagram for
sucker rods includes running a series of fatigue tests. When the
test of the first specimen is complete and the number of stress
reversals required to cause failure has been determined, an SA
identical specimen is placed in the machine and tested to failure

Smax SR
under a different load. This procedure is repeated several times
using identical specimens and a different load each time. The
results of these tests are plotted in terms of stress and number 0
of cycles to fracture. A curve (called an S-N curve) similar to the
one in Figure 17 is then drawn.
Figure 18: Variations in completely reversed cyclic stress
Fluctuating stress about a mean stress is shown:
SA = Stress Amplitude
Control (no decarburization)
80 SR = Stress Range
Maximum stress, ksi

SE = Fatigue Strength
Decarburized SM = Mean Stress
Smax = Maximum Stress
Smin = Minimum Stress
SU = Tensile Strength
104 105 106 107
Fatigue life, cycles
100 B
Figure 17: Effect of decarburization on fatigue strength of rotating
beam specimens of SAE 4140 steel, tempered for normal hardness
of Ro-48 [8]. S max
The maximum stress that will not produce failure in the 50 A ea SU
material after ten million (107) cycles is referred to as the
endurance limit. It has been determined that when steel is
SE 103 psi

tested under normal reverse bending stress conditions, its

fatigue strength (endurance limit) is approximately 50% of


the tensile strength. The fatigue strength of steel with rough 0

surface, as in hot-rolled sucker rods with no corrosion 50 100
SM 103
effects, is approximately 1/3 of the tensile strength when
tested under reverse bending.
Tensile strength = 100,000 psi

The fatigue strengths under these various test conditions can 50 Endurance limit in reverse bending = 50,000 psi

be related by a Goodman diagram to show the maximum

usable stress for sucker rod materials. It is important to note Figure 19: The Goodman diagram, based on polished samples and
that Goodman diagrams are based on a linear relationship reverse bending stress, helps to visualize the change of fatigue
with the tensile strength, not on yield strength. strength of a material under a variety of loads.

In Figure 19, the line AB represents the maximum stress with

which a material can be loaded without fatigue failure. This
stress level is at a minimum when the stress range, SR, is

1993-2016 Weatherford. All rights reserved. 13


maximum (and is a complete reversal from tension to questionable, the replacement of a worn part may be,
compression; the average stress is zero). As the stress range particularly in the absence of established standards.
is reduced (left to right on diagram), the maximum stress can
In general, wear may be defined as damage to a solid surface
be increased and the minimum stress can be increased
caused by the removal or displacement of material by the
proportionately. When the stress range is reduced to zero,
mechanical action of a contacting solid, liquid, or gas. When
the maximum stress is increased to equal the tensile
a failure is caused by one type of wear, analysis may be
strength, SU, and this is a static load.
relatively simple. However, many wear failures are caused by
The modified Goodman diagram in Figure 20 is based on combined modes of wear.
as-produced sucker rods, and the endurance limit in reverse
Sucker rods and couplings exhibit wear by one or a
bending is T/3. Because sucker rods do not operate in
combination of the following modes:
compression, the diagram is shifted to the right. Point T/4
was selected as a safety factor because it is not practical to Abrasion: Displacement of material from a surface by
operate at T/2, which is a statistical value, and scatter in test contact with hard projections on a mating surface (metal-
results is to be expected. Point T/1.75 is an arbitrary value to-metal contact) or by hard particles, such as sand and
that represents about 57% of the tensile strength and is corrosion products, trapped between two sliding surfaces
always below the yield strength. (Figure 21)
T T Adhesion wear: Wear occurs when two metallic surfaces
slide against each other under pressure (also described as
scoring, galling, seizing and scuffing)
Erosive wear: Abrasive wear involving loss of surface
material by contact with a fluid that contains foreign
i) T matter or particles
T ps
2 Stre
s 1.75 Corrosive wear: A mode of wear in which chemical or
ab electrochemical reaction contributes to the wear rate
T =A (Figure 22); for example, the pitting caused by
+ SA
3 T CO2 corrosion.
4 Erosion-Corrosion: A type of wear in which there is
relative movement between a surface and a corrosive
fluid (Figure 23). The fluid may or may not contain

abrasive particles. In cavitation erosion, the repeated

formation and collapse of vapor bubbles at the surface
imposes contact stresses that may cause pitting
or spalling.

T = SU = Tensile
Figure 20: Modified Goodman diagram based on as-produced
sucker rods for which the endurance limit in reversed bending is T/3

The Handbook on Failure Analysis and Prevention from the
American Society for Metal describes wear as a surface
phenomenon that occurs by displacement and detachment
of material. Because wear usually implies a progressive loss
of weight and alteration of dimensions over a period of time,
wear problems generally differ from those entailing
outright breakage.3
Although worn parts may break, it is more likely that a worn
part will be removed from service because it no longer
performs satisfactorily or because its performance is
marginal. Although the replacement of a broken part is not Figure 21: Extreme abrasion wear on a coupling

1993-2016 Weatherford. All rights reserved. 14


Figure 22: Corrosive wear of a rod body. The rod has rubbed against the tubing, exposed clean metal, and initiated
localized corrosion attack and subsequent pitting and grooving.

Figure 23: A rod and guide exhibiting erosion-corrosion of the rod from well fluids passing through the gap in the guide.

Tensile Failures
Typically, tensile failure of a sucker rod is not a service-oriented
failure. Such failures typically have one of these causes:
Tensile stress overload of the string while trying to free a
stuck pump
Pulling the pin off the rod upset while over-tightening the
joint with uncalibrated power tongs
Torsional overload in a PCP application
When load exceeds the tensile strength of the rod or pin, the
failure is identifiable by the necked-down area (shown in Figure 25: Tensile failure showing the cone breakface sheared at
Figure 24) and cup and cone (Figures 25 and 26) with the 45 45. This was a lab failure created during routine testing.
shear lip.

Figure Figure 26: Sucker rod pulled in two when a stuck sucker rod string
24: A photo of a sucker rod pin overtightened using power tongs. was overpulled.
Note the necking down of the pin undercut.

1993-2016 Weatherford. All rights reserved. 15


Designing to Reduce Failures

We have discussed sucker rod and sucker rod coupling

failure mechanisms and presented examples of each failure
mode. Understanding of the factors contributing to service
failures is necessary to control them. Corrective action can
improve the service life of the equipment.
This chapter will describe various conditions that cause
sucker rod and coupling failures and steps that are available
for preventing future failures. In most cases, illustrations
will be shown that represent the primary cause of each
failure mechanism.

Mill Defects
When a sucker rod fails prematurely, you will often hear the
customer say "bad steel," "faulty material," or "that string
was in service only a few months, and the string it replaced Figure 27: Surface damage caused by a sliver
lasted 5 to 6 yearsmust be a bad heat of steel." For mill defects, the obvious corrective measure is to inspect
These opinions are rarely justified because failures caused by the affected string, preferably by flux-leakage magnetic
faulty material almost never happen. However, if a customer equipment, and to discard rods with surface defects.
does experience a failure related to a mill defect, the cause Inspection service companies provide this service. Also most
will most likely be a surface defect such as a scab or sliver. manufacturers of sucker rods have inspection equipment in
house for inspection of bars before processing.
Figure 27 shows a scab (sliver), a loose or torn segment of
material or debris rolled into the surface of the bar. One end Manufacturing Problems
of these particles of metal is metallurgically bonded to the
In rare cases, sucker rods are shipped with manufacturing
body of the rod. The remaining section is rolled into the bar
defects. The most common defects are forging laps in the
surface but only attached physically. If the particle is
bead, undersize pin threads, oversize coupling threads,
dislodged, a deep surface pit remains, normally with a sharp
forged-in scale pits in the rod body adjacent to the bead,
root. The scabs and slivers act as notches in the surface of
forging laps on the square, and deep steel stamp marks in
the metal, reducing the fatigue-endurance limit by a ratio of
the flats of the upset square.
perhaps two or three to one. The condition then develops
into a notch fatigue failure. Please bear in mind that when we talk about manufacturing
defects, we are including all sucker rod manufacturers. No
manufacturer is exempt. You will find, however, that failures
caused by manufacturing defects are very uncommon.
A deep stamp mark in the steel can result in a fatigue crack
that progresses to ultimate failure. The stamp mark is a
sharp notch that raises local stress and increases surface
stresses in the notch to the point of exceeding the
endurance limit. An example can be seen in Figure 28. This
type of problem can be controlled by using steel stamps with
less sharpness to reduce penetration in the square and
reduce root sharpness at the bottom of the stamp mark.
Sometimes early failures in surface notches in the upset are
caused by a combination of high loading and the notch
effect. In such cases, redesigning the string to a lower

1993-2016 Weatherford. All rights reserved. 16


operating stress level will eliminate the failure in the square

and improve service life.

Figure 29: Deformation in the surface of a rod and the resulting

fatigue crack that led to tensile failure.

Improper Joint Makeup

For proper makeup, the API sucker rod joint is designed so
that the pin is in tension. The important factor is that the
joint must be tightened sufficiently to induce a preload in
the pina preload high enough to prevent the contact faces
from separating when the string is under its maximum
tensile load. If the joint has insufficient torque, the first full
thread root will not only be subjected to a high range of
stress, but will also be exposed to bending, as seen in
Figure 30. Early failure occurs if the rod string carries any
appreciable load.

Figure 28: Classic example of a notch fatigue failure

Handling Problems
Mishandling is a factor in rod and coupling failures. Examples
include mechanical damage to the rod surface, improper
joint makeup, bending or kinking the rods, and hammering
the surface of couplings.
In most cases, damage by mishandling could have been
avoided or the damaged part could have been discarded to
eliminate the possibility of an early failure. Training of field
personnel will reduce handling-related problems.
Field personnel can support users by keeping them informed
of any handling-related failures. Field personnel might also
suggest a course of action that could be effective in
correcting handling problems.

Mechanical Damage
Usually, mechanical damage is the result of a permanent
deformation in the surface of the rod. Figure 29 shows an
example in which fatigue started at the root of the damage
and progressed to a depth at which the rod cross-section
could not support the operating load.

Figure 30: Results of improper makeup of coupling to rod. The photo

on the top shows damage to the first full thread root; the photo on
the bottom shows damage to the coupling. Both failures were
caused by under-torquing.

1993-2016 Weatherford. All rights reserved. 17


Applying the proper circumferential displacement to the measured by dial indicator riding on the machined pin
joint during makeup is highly recommended. When power shoulder, the maximum allowance is 0.130 TIR.
tongs are out of calibration, too much torque can be applied
Most bent rods are caused by rough handling in shipment,
during makeup, and the pin can break under tensile
by improper handling while running the string in and out of
overload. Loose joints not only cause pin fatigue failures, but
the hole, or by dropping the string (Figure 32). Bending can
the separation of the pin-coupling faces also allows corrosive
also be caused by fluid pound, gas pound, or tagging bottom.
fluids to enter the coupling and initiate corrosion fatigue
Failure caused by bending is identified by these features:
failures of the coupling or pin. Refer to API Publication
RPIIBR, "Recommended Practice for Care and Handling of All fatigue cracks are on one side of the bar.
Sucker Rods," for proper methods to determine correct joint
Corrosion pits may or may not be present.
makeup and to control connection failures.
Separation of joint faces has also been attributed to The bar is visibly bent and failure started on the
unscrewing of the joint. Without the drag of friction on the concave side.
mating surfaces and with the smooth finish or rolled The break face is not perpendicular to the rod body.
threads, each stroke permits a little rotation until the
rod string separates with no apparent damage to either
pin or coupling.

Bent Rods
Straightness is important on heavily loaded rods. A bent rod
(Figure 31) will produce a high order of cyclic stress
variations and cause an early failure. Under load, any degree
of bend imposes higher tensile stresses on the inside, or
concave side, of the bend compared to the same load on a
straight rod. A string operating at its maximum loading will
be over stressed at the concave side of the bend. When the
endurance limit is exceeded, the string will fail from fatigue
that started on the inside of the bend.

Figure 32: Bent sucker rods most likely damaged by dropping the
rod string

Rods that are bent to the maximum API recommended

straightness, when under load, are stressed about 10% more
on the concave side (Figure 33).

Figure 31: A load of rods, some bent

API Specification 11B Twenty-seventh Edition, November 1,

2011, for sucker rods, Page 24, Section A.6.1, Straightness
and Surface Finishes specifies body and end straightness for
5/8-in. to 1 1/8-in.-diameter rods:
Body straightness: Within any 12 inches, the maximum
allowable bend is 0.065 inch (0.130 TIR)*.
*Total indicator run out (TIR) is the total dial gauge deflection measured at
the rod surface as the rod is rotated 360. The bend of TIR valves is twice
the amount measured by straight edge.

End straightness: As measured by supporting the rod body

at a distance of 6 inches from the rod pin shoulder and

1993-2016 Weatherford. All rights reserved. 18


P L = Length in inches As the rod guide contacts the tubing surface, it can remove
P = Load in pounds corrosion inhibitor or any protective scale that has formed.
E = Mod. of elas. 2910 6
= Out-of-straignt (in inches) This will increase the corrosion rate at the contact location.
d = Diameter (inches) of sucker rod When this occurs, it typically appears as a groove the width
A = Square area of sucker rod of the rod guide vane. It can also appear as though the rod
guide has worn through the tubing.

[Stress at Concave Area = ] (1) Rod guides will add weight to the rod string and increase the
d contact friction on the tubing. This should be taken into
L 4Ed
Stress = +
P consideration when using guides. Proper rod guide design and
L2 A placement are critical. An improperly designed rod and guide
system will shorten the run life of the sucker rod string.

X Poor Pumping Conditions

[Stress Concentration Factor = ] (2) Sucker-rod body wear and coupling wear can be indications
1.414 10 d
8 3 of poor pumping conditions:
SCF = 1 +
P PL2 + 1.767 10 7 (d4) Poor pumping speeds can induce compression loads on
Figure 33: Calculation of stress in concave area the pump downstroke and cause buckling and rod wear
(Figures 35 and 36). This condition can be corrected
Guided Sucker Rods by proper design of the rod string and reducing the
Sucker rod guides have evolved from simple metal scrapers pumping speed.
to a highly engineered thermoplastic product. With the large Deviated wells cause coupling and rod wear (Figure 37).
increase in directional and horizontal wells, rod guides are
used not only to remove paraffin from tubing and sucker Wear is also an indication of cork-screwed tubing from
rods, but also to protect and stabilize the sucker rod string. improper tension on the anchor or packer.
Rod guides can significantly increase the life of a sucker rod Fluid and gas pounding can cause rod buckling. Reduce
string by eliminating rod and tubing wear. However, there pump size or reduce pumping speed to correct fluid
are some disadvantages to using rod guides. pound. Gas interface can be reduced with a pressure
Application Issues regulator or application of a specific downhole pump and
During the pump cycle, all rod guides disrupt the fluid flow, gas separator.
some much worse than others. This causes an increase in Coupling wear can occur in a rod string with a rod rotator
fluid velocity and a low pressure zone. Because of this installed but no rod guide protection (Figure 38).
disruption, a low pressure zone is created on the upper side
(closest to the surface) of the rod guide, as seen in Figure 34.
The increase in fluid velocity can cause fluid erosion, and the
low-pressure zone can cause erosion-corrosion and CO2

Figure 35: An example of rod wear.

Figure 34: Rod guides can increase the corrosion rates on

production tubing.
Figure 36: Cracking along one side of the rod (flexing) caused by
improper pumping conditions

1993-2016 Weatherford. All rights reserved. 19


Carbon Dioxide Corrosion

Pits created by carbon dioxide corrosion are normally
deep with sharp edges and round bottoms as in Figure 39.
The scale is iron carbonate, which is hard and grey to
black in color. Pits may connect or channel in high fluid-
Figure 37: Extreme coupling and rod wear that most likely was flow environments.
caused by running through a deviated well without rod guide
protection. A spray metal coupling can be used when wear or
corrosion is a problem. Caution should be taken when using spray
metal couplings.

Figure 39: Corrosion damage to sucker rods by a carbon dioxide

(sweet corrosion) environment

Hydrogen Sulfide Corrosion

The iron sulfide produced by the action of hydrogen sulfide
Figure 38: Coupling wear in a rod string with a rod rotator installed
and water on steel typically adheres to the steel surface as a
but no rod guide protection
black powder or scale. The scale tends to cause a local
Corrosion Problems acceleration of corrosion because the iron sulfide is cathodic
to the steel. The pits are usually scattered on the metal
As discussed earlier, the word corrosion denotes destruction surface and are saucer-shaped with round edges (Figure 40).
of metal by chemical or electrochemical action. Chemical Cracks will form in the root of the pit. When placed in dilute
corrosion, although starting rapidly, often slows as soon as hydrochloric acid, the corrosion by-product (iron sulfide) will
an obstructive layer of corrosion products forms upon the release an odor like that of rotten eggs. The hydrogen
metal surface. If, however, this corrosion product is released in the reaction enters the steel to cause
continuously being cracked by bending or being removed by embrittlement or to form molecular hydrogen, which leads
rubbing or other mechanical action, corrosion will continue to blisters and cracks.
unchecked at its original rapid rate. Familiar examples of this
conjoint action on sucker rods are corrosion fatigue (in
which cyclic stress ruptures the corrosion by-product layer),
down-hole wear, and impingement attack by gas and fluid.
The major corrosives encountered in oil wells are carbon
dioxide, hydrogen sulfide, and oxygen dissolved in water.
Practically all well fluids produced by sub-surface pumps and
rods are corrosive to some degree. The corrosivity varies not
only from field to field, but also from well to well in the same
field. It also varies with time in any well.
The different types of corrosion are generally characterized
by pit shape and scale formation.

Figure 40: Corrosion damage by a hydrogen sulfide environment

(sour corrosion)

1993-2016 Weatherford. All rights reserved. 20


MIC (Microbiologically Influenced Corrosion)

MIC is a form of corrosion that originates from a large
growth of bacteria that can live in oxygen-rich (aerobic) or
oxygen-free (anaerobic) environments. Downhole
environments tend to be oxygen free, so most MIC found in
the wells are anaerobic. MIC tends to be localized, so large
isolated pits can form on sucker rods. Two types of MIC are
typically found downhole: sulfate reducing bacteria (SRB)
and acid producing bacteria (APB). Both forms can cause
significant damage if left untreated.
SRB generate hydrogen sulfide and contribute to localized Figure 43: Example of APB corrosion
corrosion by their ability to grow in the absence of oxygen. Oxygen Corrosion
The hydrogen sulfide reacts with iron in solution to form iron
sulfide (FeS) precipitate and scale. The FeS scale is cathodic to Subsurface equipment in oil wells is subject to oxygen
steel (Figure 41), SRB corrosion pits tend to be isolated and to corrosion only if oxygen from the air is introduced into the
have shallow bottoms with soft edges. Many times SRB well. The presence of carbon dioxide or hydrogen sulfide
corrosion creates worm-like pits surrounding the larger pit. increases the rate of oxygen corrosion. Corrosion of
downhole equipment in the oxygen environment usually
shows a general form of attack, sometimes producing large,
shallow, flat-bottom pits (Figure 44).

Figure 41: The FeS scale is cathodic to steel, resulting in corrosion of

scale-free areas which, in turn, adds more iron to the solution.

Acid-producing bacteria produce organic acids as a by-

product that reduces the pH, which can dissolve the sucker Figure 44: A typical form of oxygen corrosion. Localized attack may
rod. APB corrosion pits tend to interconnect with sharp- result in deep pitting, and the corrosion by-product is ferric oxide.
edged, flat-bottomed pits. Figures 42 and 43 show examples
of both types of MIC corrosion. Remedial measures
Improving corrosion-fatigue life of sucker rods requires
inhibiting corrosive wells, which also gives added protection
to the tubing string and the well casing. Because corrosion is
a surface reaction, any modification of the steel-corrosive
media interface will affect the rate of corrosion. When
added to a corrosive system, specific chemicals, called
inhibitors, modify the interface to reduce the corrosion rate.
All of the major inhibitor suppliers can furnish effective
Figure 42: Example of SRB corrosion inhibitors and proper application for reducing corrosion in
most fields. Even under the best conditions, however,
inhibitors will not be 100% effective.
Protective coatings, such as epoxy, have been used with
some success on sucker rods. The difficulty is achieving a
coating free from pin holes and handling damage. Spray
metal couplings have been used successfully for many years
to reduce corrosion rates.
Oxygen corrosion in oil wells is best controlled by the
exclusion of oxygen. The casing valve should always be
closed to the atmosphere. If production is reduced by closing
the casing, then a small check valve or ball and seat should
be installed on the casing. This will allow gas to vent to the

1993-2016 Weatherford. All rights reserved. 21


atmosphere by holding only an ounce or two of pressure to

exclude oxygen from the annulus.
Rod Type Size (in.) (lb) (DaN)
The beneficial effects of lowering maximum stress levels in a
corrosive environment have been discussed previously. 5/8 23,400 10,400

Using alloy rods in place of carbon or carbon manganese 3/4 33,800 15,000
rods has been successful as a means of improving rod life in 7/8 45,900 20,400
a corrosive environment, but no sucker rod is impervious to 1 60,000 26,600
5/8 27,600 12,200
The most reliable way to avoid sulfide stress cracking is to 3/4 39,700 17,600
use non-susceptible materials. If high-strength rods are
necessary, then an effective inhibition program should D 7/8 54,100 24,000
be used. 1 70,600 31,400

Problems Caused by a Stuck Pump 1-1/8 89,400 39,700

3/4 37,700 16,800
Pulling on the rod string to unseat a stuck pump can cause
accidental overload, excessive rod stretch, or tensile 7/8 51,400 22,800
breakage. Excessive stretch occurs when the load exceeds 1 67,100 29,800
the yield strength. Breakage occurs as the load increases 1-1/8 84,900 37,700
from beyond the yield strength to the tensile strength. If
sucker rods are permanently stretched by overload, there 3/4 45,700 20,300
will most likely be localized, external, and perhaps internal Grade HD 7/8 62,200 27,600
damage that may lead to early failures. Consequently, the T66/XD 1 81,200 36,100
affected rods should be removed from service.
1-1/8 102,800 45,700
Calculating the maximum allowable pull can prevent
3/4 43,700 19,400
stretching and tensile failure (Table 1).
7/8 59,500 26,400
S67 67D
1 77,700 34,500
1-1/8 98,400 43,700
3/4 45,700 20,300
S87 7/8 62,200 27,600
1 81,200 36,100
3/4 51,600 22,900
7/8 70,300 31,200
1 91,800 40,800
1-1/8 116,200 51,700
5/8 35,900 15,900
3/4 51,600 22,900
EL rod 7/8 70,300 31,200
1 91,800 40,800
1-1/8 116,200 51,700
Table 1: Maximum weight indicator pull (load) that can be applied
to a stuck sucker-rod string

1993-2016 Weatherford. All rights reserved. 22


Weight before rerunning. If the pump cannot be unseated, the

tubing should be pulled and rods backed off.
Size (in.) (lb/ft) (kg/m)
Hammering the Surface of Couplings
5/8 1.114 1.657 When pulling a well or when a coupling needs to be
removed after installation, many rig crews warm up the
3/4 1.634 2.432
coupling by hammering. Hammer blows cause mechanical
7/8 2.224 3.310 damage to the coupling and can induce cracking. The cracks
can be stress raisers and sites for corrosion fatigue. Damage
1 2.904 4.322 from hammering can become points where localized
corrosion will start its attack. Hammering the faces of any
1-1/8 3.676 5.471 coupling may result in an improper joint makeup that can
cause pin fatigue.
Table 2: Weight of sucker rods per foot
Hammering on spray metal couplings causes surface cracks
Note: The ratings are based on 90% of the minimum yield strength for a sucker-rod in the hard surfacing (Figure 45) and will cause localized
string in like new condition. The maximum pull should be reached with a steady
pull and not a shock load. For a tapered string, calculate the weight of the sucker corrosion and fatigue. Any coupling that has been
rod above the smallest and lowest section, and add the calculated weight to the hammered on should be discarded.
value tabulated here for the type and size of the lower section. For a single-taper
sucker-rod string, the values tabulated here are the maximum pull.

SF safety factor
Sa allowable stress (psi)
Sy yield strength of sucker rod (psi)
A cross-sectional area of sucker rod (in2)
L load in lbs
Sa = Sy x SF
L = Sa x A
weight of rods = weight per foot (see Table 2) x length of rod
Sy sucker rod maximum yield strength (psi)
Figure 45: Spray metal coupling with damage caused by hammer
Type of Rod API C blows to the surface
Minimum yield strength 60,000 PSI Thread Galling
Smallest rod 3/4 in.
Thread galling can occur because of damaged or dirty
500 feet of 7/8-in. rods above the top 3/4-in. rod threads (Figure 46). Joints seldom cross-thread because the
pin must be aligned in the coupling recess before the first
1. 60,000 psi 0.90 = 54,000 psi
threads engage. Cross-threading, however, may be possible
2. 54,000 psi 0.442 sq. in. = 23,868 lb when power tongs are used in field assembly and the
threads are damaged during stubbing.
3. Weight of 7/8-in. rods above the top 3/4-in. rod
500 ft 2.224 lb/ft = 1,112 lb Coupling and pin threads must be clean and lubricated
before assembly; if power tongs are used, the API RP11BR
4. Maximum allowable pull in pounds
recommendation for sucker rod joint makeup should
25,194 + 1,112 = 24,980 lb
be followed.
Rods that have been in service for a length of time may be
damaged by corrosion or corrosion-fatigue to a degree that
breakage will occur when the maximum calculated loads are
applied. Under these conditions, the string should be
thoroughly inspected, and affected rods should be discarded

1993-2016 Weatherford. All rights reserved. 23


Failure Analysis
In discussing failure analysis work, you must consider failure
prevention. Engineering teamwork is required to create the
kind of reliability that will give a product a superior
advantage over its competitors. However, a vital ingredient
in attaining such reliability is an adequate method for
analyzing the inevitable failures that occur during
engineering tests or during service. You must also divide
your products into two categories: (1) downhole equipment
with no threat of injury or life and (2) surface equipment for
which failure could be a threat to life or could cause injury.
All parts have a finite life. Acceptable limits are mostly
established by the user, economic situation, environment,
and competition. Remember that when a part fails, there is a
reason for failure and there are corrective measures. This
Figure 46: Thread galling during makeup. This rod has been
chapter is concerned primarily with general procedures,
overtightened, which stripped the threads and deformed the
makeup face.
techniques, and precautions used in the investigation and
analysis of metallurgical failures that occur in service.
Figure 47 is a schematic of the stages of failure analysis.

Figure 47: Stages of a failure analysis

1993-2016 Weatherford. All rights reserved. 24


Collection of Background Data and Selection Non -Destructive Testing (NDT)

Failure investigation should be directed first at gaining a NDT Method Capabilities
good understanding of the conditions under which the part
was operating. The investigator must know as much as Measures differences in radiation
possible about the manufacturing processes and service absorption
histories of the failed component and must reconstruct
insofar as possible the sequence of events leading to the Inclusions, porosity, cracks
failure. Unfortunately, in the majority of instances, the
investigator will receive a failed part with little information Uses high-frequency sonar to find surface
about its history and operating conditions. In such cases, the and subsurface defects
physical evidence must be the sole basis for the analysis. Inclusions, porosity, thickness of material,
Service history depends on how detailed and thorough the position of defects
recordkeeping was before the failure. Service history should
Dye Uses a die to penetrate open defects
include environmental details, such as normal or abnormal
loading, accidental overloads, cyclic loads, temperature, penetrate Surface cracks and porosity
temperature gradients, and corrosive environment. When
service data is sparse, the analyst must deduce service Uses a magnetic field and iron powder to
conditions, and much depends on his skill and judgment Magnetic locate surface and near-surface defects
because misleading deductions can be more harmful than particle
Surface cracks and defects
the absence of information.
Preliminary Examination Based on magnetic induction
The failed sample must be visually examined and documented Measures conductivity, magnetic
Eddy current
photographically to create a permanent record of the evidence permeability, physical dimensions, cracks,
for later reference in light of new information that may become porosity, and inclusions
available. The failed parts should be examined before any
surface cleaning to document the status of the surface Table 3: Nondestructive Tests
condition; for example, rust and scale that indicate the type of
environment in which the part was operating. The visual Mechanical Testing
examination typically will allow the investigator to identify the
mode of fracture (brittle, ductile, fatigue, etc.), points of Hardness testing, the simplest of the mechanical tests, is
initiation, and direction of propagation. Visual examination can often the most versatile tool available to the failure analyst:
be aided by low magnification tools such as a stereomicroscope. Assists in evaluating heat treatment (hardness
Photographic documentation should place particular requirements)
importance on fracture surfaces and surface defects. The
use of normal shadows can give depth to a surface, which Enables an approximation of the tensile strength of steel
makes it easier to photograph and to direct attention to Detects work hardening
important details.
Detects softening or hardening caused by overheating, by
Study of the Fracture decarburization, or by carbon or nitrogen pickup.
Where fractures are involved, the next step in preliminary
examination should be general photography of the entire Two other useful mechanical tests are tensile and impact
fractured part, including broken pieces, to record their size tests and ductile-to-brittle transition tests.
and condition and to show how the fracture is related to the Remember that the effects of size in fatigue stress-corrosion
components. Next should be a careful examination of the and hydrogen embrittlement testing are not well
fracture faces to determine the areas of prime interest and understood. However, on the basis of the limited evidence
what magnification is needed to bring out fine details. available, it appears that resistance to these failure
Photograph the fracture face at the magnification processes decreases as specimen size increases.
determined to give the most revealing details.

1993-2016 Weatherford. All rights reserved. 25


Selection and Preservation of Fracture Surfaces shown in Figure 48. The direction of crack growth is almost
always away from the tips of the chevrons. Chevron marks
Preservation of fracture surfaces is important to prevent occur because nearly all cracks are stopped at an early stage
important evidence from being destroyed. Usually in their development; as the crack front expands, the traces
protection during shipment will include wrapping samples in of the steps form chevron marks. The photograph in
cloth, cotton covering, or other suitable soft material. Avoid Figure 48 shows the fracture surface of a 5-in.-diameter, seal
contact of the fracture with chemicals, including water, that stem, tubular section. As indicated by the arrows, the
could result in corrosion damage. fracture reveals multiple crack propagation paths depicted
Sectioning by the chevron marks.
When sectioning specimens, use these techniques to protect
the fracture area:
Flame cut far away from fracture
Dry saw
Carbide cutoff with liquid coolant
Secondary Cracks
When a primary fracture has been damaged or corroded
such that most of the evidence is obliterated, it is desirable
to open any secondary cracks to expose their fracture
surfaces for examination. Secondary cracks may provide
more information than the primary fracture.

Macroscopic Examination of Fracture Surfaces

Examination of fracture surfaces at 1 to 100 diameter can Figure 48: Chevron patterns point to the multiple origins of the
be done with the unaided eye, a hand lens, or a low-power fracture.
stereoscopic microscope. Occasionally it may be
advantageous to use a scanning electron microscope at Where fracture surfaces show both flat and slant structures,
low magnifications. it may be generally concluded that the flat fracture occurred
first. Conversely, if a fracture has begun at a free surface, the
Extensive information can be obtained from examining a fracture-origin area is usually characterized by a total
fracture surface at low-power magnification. Fracture absence of slant fracture or shear lip.
surfaces may give an indication of the stress system that
produced failure. Failure in monotonic tension produces a Low-power examination of fracture surfaces often reveals
flat (square) fracture normal to the maximum tensile stress regions that have a different texture from the region of final
under plane-strain conditions and to a slant (shear) fracture failure. Fatigue, stress corrosion, and hydrogen
at about 45 if plane-stress conditions prevail. Because pure embrittlement fractures may all show these differences.
plane-strain and pure plane-stress conditions are ideal Figure 49, which shows the fracture surface at the sucker rod
situations that seldom occur in service, many fractures are body, is an excellent example of the type of information that
flat at the center but surrounded by a picture frame of slant can be obtained by macroscopic examination. The chevron
fractures. The slant fracturing occurs because conditions marks clearly indicate that the fracture origin is at the point
approximating plane strain operate at the center of the marked by the arrow. This region, unlike the rest of the
specimen but relax toward plane-stress near the surfaces. fracture, has no shear lip. The flat surface suggests that the
An example of this behavior is the familiar cup-and-cone stress causing the failure was tension parallel to the length
tensile fracture. In thin sheets or small-diameter rods, full of the rod. Note the difference in texture.
slant fractures may occur because axial stresses are relaxed
by plastic deformation that impedes plain-strain from
developing, making the crack propagation easier in a
slant direction.
Macroscopic examination can usually determine the
direction of crack growth and hence the origin of failure.
With brittle, flat fractures, determination depends largely on
the fracture surface showing "chevron marks" of the type

1993-2016 Weatherford. All rights reserved. 26


Mode of Fracture Typical Fracture Surface Characteristics

Cup and cone

Dull surface
Inclusion at the bottom of the dimple

Brittle Shiny
intergranular Grain boundary cracking

Cleavage fractures
Figure 49: Fracture surface of a sucker rod at approximately actual Flat
size, showing point of initiation (at arrow) Chevron marks, and
development of shear lips. Benchmarks

Metallographic Examination Striations (SEM)

Metallographic examination of polished and etched sections Fatigue Initiation sites
by optical microscopy is a vital part of failure investigation
to determine: Propagation area

Class of material and structure Zone of final fracture

Whether abnormalities are present Table 4: Modes of fractures and their characteristics
Heat treatment
Ductile Fracture
Corrosion, oxidation, work hardening of surfaces Ductile fractures are characterized by tearing of metal and
gross plastic deformation. Ductile tensile fractures in most
Characteristics of any cracks and their mode of
materials have a gray, fibrous appearance and are classified
on a macroscopic scale as flat-face (square) or shearface
Location of sample in respect to fracture origin. (slant) fractures.

Fracture Classifications Flat-face tensile fractures in ductile materials are produced

under plane-strain conditions (that is, in thick sections) with
Although a satisfactory classification of failures involving necking. These fractures typically occur normal to the
fractures does not exist, fractures will be classified in terms direction of loading, and some shear lip is formed at the
of their growth mechanism. Crack initiation will not be junction of the fracture surface and the part surface.
considered. Table 4 lists different modes of fractures and Microscopic examination at >100 of flat-face tensile
characteristics of each. fractures in ductile materials will reveal equiaxed dimples in
the flat-face region.
Brittle fractures
There are two types of brittle fractures: transgranular
clearage and intergranular. The transgranular facets
observed on brittle fractures are produced by clearage along
numerous parallel crystallographic planes, thus creating a
terraced fracture surface. The intergranular fractures are
grain surfaces that have been exposed by crack propagation
along grain boundaries (a rock candy fracture surface).

1993-2016 Weatherford. All rights reserved. 27


Fatigue Fracture
Fatigue fractures result from repeated or cyclic stresses,
each of which may be substantially below the nominal yield
strength. The most noticeable macroscopic features of
classic fatigue-fracture surfaces are the progression marks
(also known as bench marks, clamshell marks, or tide marks).
Little macroscopic ductility is associated with fatigue
fracture. Microscopically, surfaces of fatigue fractures are
characterized by presence of striations, each of which is
produced by a single cycle of stress; however, every cycle
does not produce a striation.
Stress-Corrosion Cracking
Stress-corrosion cracking is a mechanical-environmental
failure process in which mechanical stress and chemical
attack combine in the initiation and propagation of fracture
in a metal part. Stress-corrosion cracks may be intergranular,
Figure 51: Transcrystalline stress corrosion cracking in a Type 316L
transgranular, or a combination of both. Transgranular stainless steel gas lift valve assembly at 50X magnification; solution
fractures that show branching are typical of stress-corrosion etched using Frys reagent
cracking of austenitic stainless steels of an 18cr-8 ni
type. Figures 50 and 51 show examples of stress- Embrittlement by Liquid Metals
corrosion cracking. Embrittlement can occur as a result of liquid metal around
grain boundaries, that is, penetration of copper alloys by
mercury and penetration of steel by molten tine and
cadmium. Polished samples or polished and etched sections
can be examined by microscope. Positive identification is
possible by electron-microprobe analysis.
Hydrogen embrittlement (sulfide cracking)
Positive identification of hydrogen embrittlement is often
difficult, and it is frequently impossible to differentiate
between hydrogen-induced, delayed fracture and stress-
corrosion-cracking fracture. Analysis should include
the following:
Check hardness
Check for H2S
Run metallographic examination
Figure 50: Intercrystalline stress corrosion cracks in precipitation
hardened stainless steel UNS S17400 alloy at 300X magnification; Chemical Analysis
solution etched using Vilellas reagent Perform chemical analysis to confirm analysis of materials.
Unreported gaseous elements (or interstitial) have profound
effects on mechanical properties of some metals. In steel the
effects of oxygen, nitrogen, and hydrogen are of major
importance. Oxygen and nitrogen may cause strain aging
and quench aging. Hydrogen may induce embrittlement.
Failures Resulting From Corrosion Damage
Various forms of corrosion attack render parts inoperable.
These modes were discussed earlier in this document.

1993-2016 Weatherford. All rights reserved. 28



1. Bucaram, S.M., Byors, H.G., and Kaplan, M. Proper Selection, Handling, and Protection of Downhole Materials A
Practical and Economical Approach. Paper No. 55. Corrosion/72
2. Madayag. 1969. Metal Fatigue, Theory and Design. Wiley. New York. 2126.
3. Failure Analysis and Prevention. Metals Handbook. Vol. 10. American Society for Metals. Metals Park, Ohio.
4. Corrosion of Oil-and-Gas-Well Equipment. Book 2 of the Vocational Training Series. API.
5. Dvoracek, L. M. "Corrosion Fatigue Testing of Oil Well Sucker Rods Steels." Union Oil Company of California. Union
Research Center, Brea, California 92621.
6. Mehdizadeh, P., McGlasson, R. L., and Landers, J. E. 1963. "Corrosion Fatigue Performance of a Carbon Steel in
Production Environments." Paper presented to the South Central Region Conference of NACE. New Orleans, Louisiana.
October 1822.

1993-2016 Weatherford. All rights reserved. 29
Weatherford products and services are subject to the Companys standard terms and conditions, available on request or at For more information contact an authorized Weatherford
representative. Unless noted otherwise, trademarks and service marks herein are the property of Weatherford and may be registered in the United States and/or other countries. Weatherford products named
herein may be protected by one or more U.S. and/or foreign patents. Specifications are subject to change without notice. Weatherford sells its products and services in accordance with the terms and conditions
set forth in the applicable contract between Weatherford and the client.

El is a registered trademark of Weatherford in the United States. 1993-2016 Weatherford. All rights reserved. 02.2016